---

Contents

• Quick Reference
• Detailed Reference

---

Quick Reference

NAME

tablelist::tablelist – Create and manipulate tablelist widgets

SYNOPSIS

tablelist::tablelist pathName ?options?

STANDARD OPTIONS

-borderwidth          -highlightthickness  -setgrid
-cursor               -relief              -xscrollcommand
-exportselection      -selectbackground    -yscrollcommand
-highlightbackground  -selectborderwidth
-highlightcolor       -selectforeground

OPTIONS FOR THE BODY COMPONENT OF THE WIDGET

-background  -disabledforeground  -font  -foreground

WIDGET-SPECIFIC OPTIONS

-acceptchildcommand command

-acceptdropcommand command

-activestyle frame|none|underline

-arrowcolor color

-arrowdisabledcolor color

-arrowstyle flat5x3|flat5x4|flat6x4|flat7x4|flat7x5| flat7x7|flat8x5|flat9x5|flat9x6|
            flatAngle7x4| flatAngle7x5|flatAngle9x5|flatAngle9x6|flat1Angle9x7|
            flatAngle10x6| flatAngle10x7|photo7x7|sunken8x7|sunken10x9| sunken12x11

-autoscan boolean

-collapsecommand command

-columns {width title ?left|right|center? width title ?left|right|center? ...}

-columntitles {title title ...}

-customdragsource boolean

-editendcommand command

-editselectedonly boolean

-editstartcommand command

-expandcommand command

-forceeditendcommand boolean

-fullseparators boolean

-height lines

-incrarrowtype up|down

-instanttoggle boolean

-labelactivebackground color

-labelactiveforeground color

-labelbackground color  or  -labelbg color

-labelborderwidth screenDistance  or  -labelbd screenDistance

-labelcommand command

-labelcommand2 command

-labeldisabledforeground color

-labelfont font

-labelforeground color  or  -labelfg color

-labelheight lines

-labelpady screenDistance

-labelrelief raised|sunken|flat|ridge|solid| groove

-listvariable variable

-movablecolumns boolean

-movablerows boolean

-movecolumncursor cursor

-movecursor cursor

-populatecommand command

-protecttitlecolumns boolean

-resizablecolumns boolean

-resizecursor cursor

-selectmode single|browse|multiple|extended

-selecttype row|cell

-setfocus boolean

-showarrow boolean

-showeditcursor boolean

-showhorizseparator boolean

-showlabels boolean

-showseparators boolean

-snipstring string

-sortcommand command

-spacing screenDistance

-state normal|disabled

-stretch all|columnIndexList

-stripebackground color  or  -stripebg color

-stripeforeground color  or  -stripefg color

-stripeheight lines

-takefocus 0|1|""|command

-targetcolor color

-tight boolean

-titlecolumns number

-tooltipaddcommand command

-tooltipdelcommand command

-treecolumn columnIndex

-treestyle adwaita|ambiance|aqua|baghira|dust| dustSand|gtk|
           klearlooks| mate|mint|newWave|oxygen1|oxygen2|
           phase| plastik|plastique|radiance|ubuntu|
           vistaAero| vistaClassic|win7Aero|win7Classic|
           winnative| winxpBlue|winxpOlive|winxpSilver

-width characters

COLUMN CONFIGURATION OPTIONS

-align left|right|center

-background color  or  -bg color

-changesnipside boolean

-changetitlesnipside boolean

-editable boolean

-editwindow name

-font font

-foreground color  or  -fg color

-formatcommand command

-hide boolean

-labelalign left|right|center

-labelbackground color  or  -labelbg color

-labelborderwidth screenDistance  or  -labelbd screenDistance

-labelcommand command

-labelcommand2 command

-labelfont font

-labelforeground color  or  -labelfg color

-labelheight lines

-labelimage image

-labelpady screenDistance

-labelrelief raised|sunken|flat|ridge|solid| groove

-maxwidth width

-name name

-resizable boolean

-selectbackground color

-selectforeground color

-showarrow boolean

-showlinenumbers boolean

-sortcommand command

-sortmode ascii|asciinocase|command|dictionary| integer|real

-stretchable boolean

-stripebackground color

-stripeforeground color

-text list

-title title

-valign center|top|bottom

-width width

-wrap boolean

ROW CONFIGURATION OPTIONS

-background color  or  -bg color

-font font

-foreground color  or  -fg color

-hide boolean

-name name

-selectable boolean

-selectbackground color

-selectforeground color

-text list

CELL CONFIGURATION OPTIONS

-background color  or  -bg color

-editable boolean

-editwindow name

-font font

-foreground color  or  -fg color

-image image

-selectbackground color

-selectforeground color

-stretchwindow boolean

-text text

-valign center|top|bottom

-window command

-windowdestroy command

-windowupdate command

DESCRIPTION

COLORS AND FONTS

USING A TABLELIST AS MULTI-COLUMN TREE WIDGET

INTERACTIVE CELL EDITING

DRAG & DROP SUPPORT

VIRTUAL EVENTS

ROW INDICES

number  knumber  active  anchor  end  last  top  bottom  @x,y  name

NODE INDICES

root  number  knumber  active  anchor  end  last  top  bottom  @x,y  name

COLUMN INDICES

number  active  anchor  end  last  left  right  @x,y  name

CELL INDICES

row,col  active  anchor  end  last  @x,y

    row: number  knumber  active  anchor  end  last  top  bottom  name
    col: number  active  anchor  end  last  left  right  name

WIDGET COMMAND

pathName activate index

pathName activatecell cellIndex

pathName applysorting itemList

pathName attrib ?name? ?value name value ...?

pathName bbox index

pathName bodypath

pathName bodytag

pathName canceledediting

pathName cancelediting

pathName cellattrib cellIndex ?name? ?value name value ...?

pathName cellbbox cellIndex

pathName cellcget cellIndex option

pathName cellconfigure cellIndex ?option? ?value option value ...?

pathName cellindex cellIndex

pathName cellselection option args

pathName cellselection anchor cellIndex

pathName cellselection clear firstCell lastCell

pathName cellselection clear cellIndexList

pathName cellselection includes cellIndex

pathName cellselection set firstCell lastCell

pathName cellselection set cellIndexList

pathName cget option

pathName childcount nodeIndex

pathName childindex index

pathName childkeys nodeIndex

pathName collapse index ?-fully|-partly?

pathName collapseall ?-fully|-partly?

pathName columnattrib columnIndex ?name? ?value name value ...?

pathName columncget columnIndex option

pathName columnconfigure columnIndex ?option? ?value option value ...?

pathName columncount

pathName columnindex columnIndex

pathName columnwidth columnIndex ?-requested|-stretched|-total?

pathName configcelllist {cellIndex option value cellIndex option value ...}

pathName configcells ?cellIndex option value cellIndex option value ...?

pathName configcolumnlist {columnIndex option value columnIndex option value ...}

pathName configcolumns ?columnIndex option value columnIndex option value ...?

pathName configrowlist {index option value index option value ...}

pathName configrows ?index option value index option value ...?

pathName configure ?option? ?value option value ...?

pathName containing y

pathName containingcell x y

pathName containingcolumn x

pathName cornerlabelpath

pathName cornerpath

pathName curcellselection

pathName curselection

pathName delete first last

pathName delete indexList

pathName deletecolumns firstColumn lastColumn

pathName deletecolumns columnIndexList

pathName depth nodeIndex

pathName descendantcount nodeIndex

pathName editcell cellIndex

pathName editinfo

pathName editwinpath

pathName editwintag

pathName entrypath

pathName expand index ?-fully|-partly?

pathName expandall ?-fully|-partly?

pathName expandedkeys

pathName fillcolumn columnIndex text

pathName findcolumnname name

pathName findrowname name ?-descend? ?-parent nodeIndex?

pathName finishediting

pathName formatinfo

pathName get first last

pathName get indexList

pathName getcells firstCell lastCell

pathName getcells cellIndexList

pathName getcolumns firstColumn lastColumn

pathName getcolumns columnIndexList

pathName getformatted first last

pathName getformatted indexList

pathName getformattedcells firstCell lastCell

pathName getformattedcells cellIndexList

pathName getformattedcolumns firstColumn lastColumn

pathName getformattedcolumns columnIndexList

pathName getfullkeys first last

pathName getfullkeys indexList

pathName getkeys first last

pathName getkeys indexList

pathName hasattrib name

pathName hascellattrib cellIndex name

pathName hascolumnattrib columnIndex name

pathName hasrowattrib index name

pathName hidetargetmark

pathName imagelabelpath cellIndex

pathName index index

pathName insert index ?item item ...?

pathName insertchildlist parentNodeIndex childIndex itemList

pathName insertchild(ren) parentNodeIndex childIndex ?item item ...?

pathName insertcolumnlist columnIndex {width title ?left|right|center? width title ?left|right|center? ...}

pathName insertcolumns columnIndex ?width title ?left|right|center? width title ?left|right|center? ...?

pathName insertlist index itemList

pathName iselemsnipped cellIndex fullTextName

pathName isexpanded index

pathName istitlesnipped columnIndex fullTextName

pathName isviewable index

pathName itemlistvar

pathName labelpath columnIndex

pathName labels

pathName labeltag

pathName move sourceIndex targetIndex

pathName move sourceIndex targetParentNodeIndex targetChildIndex

pathName movecolumn sourceColumn targetColumn

pathName nearest y

pathName nearestcell x y

pathName nearestcolumn x

pathName noderow parentNodeIndex childIndex

pathName parentkey nodeIndex

pathName refreshsorting ?parentNodeIndex?

pathName rejectinput

pathName resetsortinfo

pathName rowattrib index ?name? ?value name value ...?

pathName rowcget index option

pathName rowconfigure index ?option? ?value option value ...?

pathName scan mark|dragto x y

pathName searchcolumn columnIndex pattern ?options?

pathName see index

pathName seecell cellIndex

pathName seecolumn columnIndex

pathName selection option args

pathName selection anchor index

pathName selection clear first last

pathName selection clear indexList

pathName selection includes index

pathName selection set first last

pathName selection set indexList

pathName separatorpath ?columnIndex?

pathName separators

pathName showtargetmark before|inside index

pathName size

pathName sort ?-increasing|-decreasing?

pathName sortbycolumn columnIndex ?-increasing|-decreasing?

pathName sortbycolumnlist columnIndexList ?sortOrderList?

pathName sortcolumn

pathName sortcolumnlist

pathName sortorder

pathName sortorderlist

pathName targetmarkpath

pathName targetmarkpos y ?-any|-horizontal|-vertical?

pathName togglecolumnhide firstColumn lastColumn

pathName togglecolumnhide columnIndexList

pathName togglerowhide first last

pathName togglerowhide indexList

pathName toplevelkey index

pathName unsetattrib name

pathName unsetcellattrib cellIndex name

pathName unsetcolumnattrib columnIndex name

pathName unsetrowattrib index name

pathName viewablerowcount ?first last?

pathName windowpath cellIndex

pathName xview args

pathName xview

pathName xview units

pathName xview moveto fraction

pathName xview scroll number units|pages

pathName yview args

pathName yview

pathName yview units

pathName yview moveto fraction

pathName yview scroll number units|pages

DEFAULT AND INDIVIDUAL BINDINGS FOR THE TABLELIST BODY

DEFAULT AND INDIVIDUAL BINDINGS FOR THE HEADER LABELS

DEFAULT BINDINGS FOR INTERACTIVE CELL EDITING

KEYWORDS

tablelist, multi-column, listbox, tree, widget

---

Detailed Reference

NAME

tablelist::tablelist – Create and manipulate tablelist widgets

SYNOPSIS

tablelist::tablelist pathName ?options?

STANDARD OPTIONS

-borderwidth          -highlightthickness  -setgrid
-cursor               -relief              -xscrollcommand
-exportselection      -selectbackground    -yscrollcommand
-highlightbackground  -selectborderwidth
-highlightcolor       -selectforeground

See the options manual entry for details on the standard options.  The -highlightbackground, -highlightcolor, and -highlightthickness options are only supported by the Tablelist package, but not by Tablelist_tile.  When using the package Tablelist_tile, the options -selectbackground, -selectborderwidth, and -selectforeground have theme-specific default values.

OPTIONS FOR THE BODY COMPONENT OF THE WIDGET

-background  -disabledforeground  -font  -foreground

These options (described in the options manual entry) are only valid for the body component of the tablelist widget.  As discussed in the next section, the colors and font used when drawing the header labels can be different from those specified for the body.  When using the package Tablelist_tile, these options have theme-specific default values.

WIDGET-SPECIFIC OPTIONS

Command-Line Name:	-acceptchildcommand
Database Name:	acceptChildCommand
Database Class:	AcceptChildCommand

Specifies a Tcl command used to decide whether a given tablelist node may accept a specified item being moved interactively as a child.  The specified command is automatically concatenated with the name of the tablelist widget, the node index of the would-be new parent node, and the row index of the dragged item, the resulting script is evaluated in the global scope, and the return value (which must be a boolean) will determine whether to allow to move the source item to the current mouse position as a child of the given parent node.

For example, in the case of a tablelist widget used as a file manager, in which the top-level items represent volumes mounted on the system, the command specified by this option might look like in the (pseudo-)code below:

proc acceptChildCmd {tbl targetParentNodeIdx sourceRow} {
    if {[string compare $targetParentNodeIdx "root"] == 0} {
        # Allow only volumes as top-level items
        return [expr {[$tbl depth $sourceRow] == 1}]
    } else {
        # Allow only directories as parent items
        return [$targetParentNodeIdx represents a directory]
    }
}

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this option is not supported for Tk versions earlier than 8.3.

Command-Line Name:	-acceptdropcommand
Database Name:	acceptDropCommand
Database Class:	AcceptDropCommand

Specifies a Tcl command used to decide whether a given tablelist row may accept a specified item being moved interactively to be dropped as a sibling just before that row.  The specified command is automatically concatenated with the name of the tablelist widget, the row index of the would-be new target, and the row index of the dragged item, the resulting script is evaluated in the global scope, and the return value (which must be a boolean) will determine whether to allow to move the source item to the current mouse position, just before the given row.

For example, to make sure that the position of the last row of a tablelist widget won't be affected by the interactive row move operation, the command specified by this option might look like in the code below:

proc acceptDropCmd {tbl targetRow sourceRow} {
    set rowCount [[$tbl size]
    return [expr {$sourceRow != $rowCount - 1 && $targetRow < $rowCount}]
}

Command-Line Name:	-activestyle
Database Name:	activeStyle
Database Class:	ActiveStyle

Specifies how to diplay the active item or element (depending on the value of the -selecttype configuration option) when the tablelist has the keyboard focus.  The allowed values are frame, none, and underline.  The default value frame shows a thin frame around the active item or element, which in most cases looks nice.  It looks less pretty when applied to the active item if the background color of some of its cells was changed by using the columnconfigure or cellconfigure widget command and no column separators are shown.  The value none specifies that no special indication of the active item or element is to be performed.  The value underline produces the same visual effect as in the case of the Tk core listbox.

Command-Line Name:	-arrowcolor
Database Name:	arrowColor
Database Class:	ArrowColor

Specifies the color to use for the up- or down-arrow placed into a column label by the sortbycolumn or sortbycolumnlist subcommand of the Tcl command associated with the widget.  This option is only relevant if the value of the -showarrow option is true.  The default value depends on the windowing system in the Tablelist package and on the current theme in Tablelist_tile.  For example, if the windowing system is x11 then the default is an empty string, indicating that the arrow will inherit the background color of the label in which it is placed (but is distinguishable from the latter, due to its 3-D border and sunken relief, because in this case the -arrowstyle option has the default value sunken10x9).  On the windowing system win32, the default arrow color is #aca899 for Windows XP, #569bc0 for Windows Vista and Windows 7, and an empty string for older Windows versions, paired with the default arrow style flat9x5, flat7x4, and sunken8x7, respectively.  Finally, for the windowing systems classic and aqua on the Macintosh, the default arrow color and style depend on the version of the operating system:  Starting with version 10.10 (Yosemite) of Mac OS X, the default arrow color for the windowing system aqua is #404040 and the default arrow style is flatAngle7x4.  For earlier operating system versions, the default values are the arrow color #717171 and the arrow style flat7x7 (or photo7x7, in the presence of PNG support).

Command-Line Name:	-arrowdisabledcolor
Database Name:	arrowDisabledColor
Database Class:	ArrowDisabledColor

Specifies the color to use for the up- or down-arrow placed into a column label by the sortbycolumn or sortbycolumnlist subcommand of the Tcl command associated with the widget when the tablelist's state is disabled.  This option is only relevant if the value of the -showarrow option is true.  When the default value of the -arrowcolor option is an empty string then this is the default for the -arrowdisabledcolor option, too; otherwise the latter's default value equals the default foreground color of the header labels in disabled state.

Command-Line Name:	-arrowstyle
Database Name:	arrowStyle
Database Class:	ArrowStyle

Specifies the relief, shape, width, and height of the up- or down-arrow placed into a column label by the sortbycolumn or sortbycolumnlist subcommand of the Tcl command associated with the widget.  This option is only relevant if the value of the -showarrow option is true.  The currently supported values are flat5x3, flat5x4, flat6x4, flat7x4, flat7x5, flat7x7, flat8x5, flat9x5, flat9x6, flatAngle7x4, flatAngle7x5, flatAngle9x5, flatAngle9x6, flatAngle9x7, flatAngle10x6, flatAngle10x7, photo7x7, sunken8x7, sunken10x9, and sunken12x11, as shown in the picture below.  The default value depends on the windowing system in the Tablelist package and on the current theme in Tablelist_tile; see the description of the -arrowcolor option for details.

While the sort arrows of the styles flat* and sunken* are created from bitmaps, the arrow style photo7x7 uses PNG images that look and behave very close to the native sort arrows on Mac OS X Aqua versions earlier than 10.10 (Yosemite).  This arrow style is only supported if the Tk version is either 8.6 (with built-in PNG support), or 8.5 and the img::png package can be loaded into the interpreter.  When supported, photo7x7 will be the default value of the -arrowstyle option in Tablelist_tile with the aqua theme for Mac OS X versions earlier than 10.10 (otherwise flat7x7 will be used as default for these earlier OS X versions).  When using this arrow style, the -arrowcolor and -arrowdisabledcolor options have no effect, but, due to the transparency information contained in the PNG images, the arrows will automatically adapt their color to the various states of the header labels, just like the native sort arrows.

Command-Line Name:	-autoscan
Database Name:	autoScan
Database Class:	AutoScan

Specifies a boolean value that controls whether to trigger the automatic scrolling when the mouse leaves the tablelist window with button 1 down.  The default is 1, meaning that automatic scrolling will be in effect, just like in the case of the Tk listbox widget.  However, when using the TkDND package or some other drag & drop implementation, you might want to set this option to 0, in order to avoid any conflicts between the drag operation and the automatic scrolling.

REMARK:  Starting with Tablelist version 5.12, this option has become obsolete, because the default bindings now suppress the above-mentioned automatic scrolling if the tablelist's body component was registered as a drag source for mouse button 1 via the  tkdnd::drag_source register  or the BWidget DragSite::register command, or the tablelist's -customdragsource option was set to true.

Command-Line Name:	-collapsecommand
Database Name:	collapseCommand
Database Class:	CollapseCommand

Specifies a Tcl command to be invoked when collapsing a row of a tablelist used as a tree widget (with the aid of the collapse or collapseall subcommand).  The specified command is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope, before hiding the descendants of the row in question.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this option is not supported for Tk versions earlier than 8.3.

Command-Line Name:	-columns
Database Name:	columns
Database Class:	Columns

Specifies the widths, titles, and alignments of the columns.  The option's value must be a list of the form

width title ?alignment? width title ?alignment? ...

Each width must be a number.  A positive value specifies the column's width in average-size characters of the widget's font.  If width is negative, its absolute value is interpreted as a column width in pixels.  Finally, a value of zero specifies that the column's width is to be made just large enough to hold all the elements in the column, including its header (but no larger than the maximum width indicated by the -maxwidth column configuration option).  In all three cases, the effective column width will be somewhat greater because of the margins created automatically to the left and right of the column.

Each title specifies the text to be displayed in the column's header, and may optionally be followed in the next list element by an alignment, which specifies how to align the elements of the column.  Each alignment must be one of left, right, or center.  The default is left.  The alignment also refers to the column's title as long as the -labelalign option hasn't been specified for that column, or if its value is an empty string.

The default value of this option is an empty list, specifying that initially the widget has no columns.

REMARK:  Columns whose width was specified as zero are called dynamic-width columns.  In general, they are more user-friendly than their static-width counterparts, being that their widths are automatically adapted to their contents.  On the other hand, the static-width columns perform significantly better on item insertion and sorting than the dynamic-width ones, due to some optimizations introduced in Tablelist version 5.6.  (Prior to that release, the above-mentioned operations were faster with dynamic-width columns than with static-width ones.)

Command-Line Name:	-columntitles
Database Name:	columnTitles
Database Class:	ColumnTitles

This option provides a simplified form of specifying dynamic-width, left-aligned tablelist columns.  Its value is viewed as a list of column titles.  The default is an empty list.

In the simplest case that no columns have been specified yet, setting this option to the value given by the list

title title ...

is equivalent to setting the -columns option to the value given by the list

0 title left 0 title left ...

If the columns have already been specified then this option updates their titles (as many of them as possible) and, if the number of elements of its value is greater than the number of columns then it uses the remaining elements as titles of additional dynamic-width, left-aligned columns.  For example, if the widget has 3 columns and the option's value is a list of length 5 then the option will update the titles of the 3 columns and will append 2 new dynamic-width, left-aligned columns having as titles the last 2 elements of the list.  If the widget has 3 columns and the option specifies just 2 texts then it will update the titles of the first 2 columns only.

Command-Line Name:	-customdragsource
Database Name:	customDragSource
Database Class:	CustomDragSource

Specifies a boolean value that indicates whether the tablelist widget is a drag source for some drag & drop implementation other than TkDND and the drag & drop framework included in BWidget.  If true then the default bindings will perform an automatic drag-friendly handling of the selection and will suppress the automatic scrolling when the mouse leaves the tablelist window with button 1 down.  The default is 0.

Command-Line Name:	-editendcommand
Database Name:	editEndCommand
Database Class:	EditEndCommand

Specifies a Tcl command to be invoked on normal termination of the interactive editing of a cell's contents if the final text of the temporary embedded widget used for the editing is different from its initial one.  The command is automatically concatenated with the name of the tablelist widget, the cell's row and column indices, as well as the final contents of the edit window, the resulting script is evaluated in the global scope, and the return value becomes the cell's new contents after destroying the temporary embedded widget.  The main purpose of this script is to perform a final validation of the edit window's contents.  See the description of the -forceeditendcommand option for more about the invocation of the command mentioned above, as well as the INTERACTIVE CELL EDITING section for details on the editing process.

Command-Line Name:	-editselectedonly
Database Name:	editSelectedOnly
Database Class:	EditSelectedOnly

Specifies a boolean value that controls whether to start the interactive cell editing when mouse button 1 is pressed in an editable cell.  If this value is true then the editing will only be started if the cell has previously been selected (interactively or programmatically).  In this case, a first left-click will usually just select the cell (or its row, depending on the value of the -selecttype option), and a second mouse click will start the editing session.  The default is 0, meaning that the editing will be started regardless of whether the cell is selected or not.

Command-Line Name:	-editstartcommand
Database Name:	editStartCommand
Database Class:	EditStartCommand

Specifies a Tcl command to be invoked when the interactive editing of a cell's contents is started.  The command is automatically concatenated with the name of the tablelist widget, the cell's row and column indices, as well as the text displayed in the cell, the resulting script is evaluated in the global scope, and the return value becomes the initial contents of the temporary embedded widget used for the editing.  The main purpose of this script is to define validations for the edit window's contents.  See the INTERACTIVE CELL EDITING section for details on the editing process.

Command-Line Name:	-expandcommand
Database Name:	expandCommand
Database Class:	ExpandCommand

Specifies a Tcl command to be invoked when expanding a row of a tablelist used as a tree widget (with the aid of the expand or expandall subcommand).  The specified command is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope, before displaying the children of the row in question.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this option is not supported for Tk versions earlier than 8.3.

REMARK:  It is common practice to use the command specified as the value of this option to insert the children of the row that is about to be expanded, if it has no children yet.  For example, the Directory Viewer demo script uses the command implemented as follows:

proc expandCmd {tbl row} {
    if {[$tbl childcount $row] == 0} {
        # Get the name of the directory whose leaf name is
        # displayed in the first cell of the specified row
        set dir [$tbl rowattrib $row pathName]

        # Display the contents of the directory $dir
        # as child items of the one identified by $row
        putContents $dir $tbl $row
    }

    # The rest is just eye candy:
    if {[$tbl childcount $row] != 0} {
        # Update the image displayed in the row's first cell
        $tbl cellconfigure $row,0 -image openFolderImg
    }
}

Command-Line Name:	-forceeditendcommand
Database Name:	forceEditEndCommand
Database Class:	ForceEditEndCommand

Specifies a boolean value that controls the invocation of the command given by the the -editendcommand option.  If this value is true then the command will be invoked on normal termination of the editing process even if the final text of the temporary embedded widget used for the editing equals its initial one, and will also be invoked when the interactive cell editing is canceled (in the latter case, the text passed to it as last argument will be the cell's original contents, not its final one).  The default value of this option is 0, meaning that the command will only be invoked on normal termination of the editing process, if the final text of the temporary embedded widget is different from its initial one.  See the INTERACTIVE CELL EDITING section for details on the editing process.

If the option's value is true and no value for the -editendcommand option was specified, then on normal termination of the editing process the cell's new contents will be set to the text contained in the edit window, even if it has not been changed interactively (but might have been returned by the command given by the the -editstartcommand option).

Setting this option to true enables you to execute an arbitrary action whenever the interactive cell editing is finished.  Just binding a script to the <Destroy> event for the temporary embedded widget used for the editing won't work, because that widget might be destroyed and recreated automatically under various circumstances.  Alternately, you can use the <<TablelistCellUpdated>> and <<TablelistCellRestored>> virtual events, generated by the finishediting and cancelediting subcommands, respectively.

Command-Line Name:	-fullseparators
Database Name:	fullSeparators
Database Class:	FullSeparators

Specifies a boolean value that controls whether the separators (if any) shall extend all of the way to the bottom of the tablelist's body.  The default is 0, meaning that the height of the separators will be adjusted to the widget's content, i.e., they won't extend to the bottom of the tablelist's body if there is free vertical space left below the widget's last row.

Command-Line Name:	-height
Database Name:	height
Database Class:	Height

Specifies the desired height for the window, in units of characters in the font given by the -font option.  This is at the same time the desired height in lines, provided that no column-, row-, or cell-specific fonts, multi-line elements, or embedded images or windows will make the height of any of the rows different from the one corresponding to the above-mentioned font.  With this restriction, if the option's value is zero or less, then the window's height is made just large enough to hold all the items in the tablelist widget.

Command-Line Name:	-incrarrowtype
Database Name:	incrArrowType
Database Class:	IncrArrowType

Specifies the type of the arrow placed into a column label when sorting the items based on that column in increasing order, with the aid of the sortbycolumn or sortbycolumnlist subcommand of the Tcl command associated with the widget.  The value of this option must be one of up or down.  The default is up.  This option is only relevant if the value of the -showarrow option is true.

Command-Line Name:	-instanttoggle
Database Name:	instantToggle
Database Class:	InstantToggle

Specifies a boolean value that controls whether the interactive cell editing with the aid of a Tk or tile checkbutton widget, if started with the left mouse button, will be finished immediately after automatically toggling the checked state of the temporary embedded checkbutton.  The default value is 0, meaning that a mouse click into the cell will only start the editing session (and automatically toggle the checkbutton's checked state), but not also finish it.

Command-Line Name:	-labelactivebackground
Database Name:	labelActiveBackground
Database Class:	Foreground

Specifies the -activebackground option for the header labels, i.e., the background color to use when the mouse cursor is positioned over a header label and the value of tk_strictMotif is false.  This option is only defined in the Tablelist package if the Tk version being used supports the -activebackground option for label widgets.  This is checked by Tablelist at initialization time, and will normally be the case for Tk versions 8.3.2 or higher.  On the other hand, the Tablelist_tile package doesn't support the -labelactivebackground option.

Command-Line Name:	-labelactiveforeground
Database Name:	labelActiveForeground
Database Class:	Background

Specifies the -activeforeground option for the header labels, i.e., the foreground color to use when the mouse cursor is positioned over a header label and the value of tk_strictMotif is false.  This option is only defined in the Tablelist package if the Tk version being used supports the -activeforeground option for label widgets.  This is checked by Tablelist at initialization time, and will normally be the case for Tk versions 8.3.2 or higher.  On the other hand, the Tablelist_tile package doesn't support the -labelactiveforeground option.

Command-Line Name:	-labelbackground or -labelbg
Database Name:	labelBackground
Database Class:	Background

Specifies the -background option for the header labels.  This option is only supported by the Tablelist package, but not by Tablelist_tile.

Command-Line Name:	-labelborderwidth or -labelbd
Database Name:	labelBorderWidth
Database Class:	BorderWidth

Specifies the -borderwidth option for the header labels.  This option is different from the standard -borderwidth option defined for the tablelist widget itself.  In the package Tablelist_tile this option has a theme-specific default value.

REMARK:  Please take into account that in some themes, setting this option to a value other than the default might be ignored by tile and thus could cause alignment problems.  This is because the border of tile widgets is drawn with theme-specific methods, which will not always produce the results known from Tk widgets.

Command-Line Name:	-labelcommand
Database Name:	labelCommand
Database Class:	LabelCommand

Specifies the Tcl command to be invoked when mouse button 1 is pressed over one of the header labels and later released over the same label.  When the <ButtonRelease-1> event occurs, the command is automatically concatenated with the name of the tablelist widget and the column index of the respective label, and the resulting script is evaluated in the global scope.  If the tablelist's state is disabled then this action will not take place.  The most common value of this option is tablelist::sortByColumn; this command sorts the items based on the column whose index was passed to it as second argument.

Command-Line Name:	-labelcommand2
Database Name:	labelCommand2
Database Class:	LabelCommand2

Specifies the Tcl command to be invoked when mouse button 1 is pressed together with the Shift key over one of the header labels and later released over the same label.  When the <ButtonRelease-1> event occurs, the command is automatically concatenated with the name of the tablelist widget and the column index of the respective label, and the resulting script is evaluated in the global scope.  If the tablelist's state is disabled then this action will not take place.  The most common value of this option is tablelist::addToSortColumns; this command adds the column index passed to it as second argument to the list of sort columns and sorts the items based on the columns indicated by the modified list.

Command-Line Name:	-labeldisabledforeground
Database Name:	labelDisabledForeground
Database Class:	DisabledForeground

Specifies the -disabledforeground option for the header labels, i.e., the foreground color to use for the labels when the tablelist's state is disabled.  This option is only defined in the Tablelist package if the Tk version being used supports the -disabledforeground option for label widgets.  This is checked by Tablelist at initialization time, and will normally be the case for Tk versions 8.3.1 or higher.  On the other hand, the Tablelist_tile package doesn't support the -labeldisabledforeground option.

Command-Line Name:	-labelfont
Database Name:	labelFont
Database Class:	Font

Specifies the -font option for the header labels.  In the package Tablelist_tile this option has a theme-specific default value.

Command-Line Name:	-labelforeground or -labelfg
Database Name:	labelForeground
Database Class:	Foreground

Specifies the -foreground option for the header labels.  In the package Tablelist_tile this option has a theme-specific default value.

Command-Line Name:	-labelheight
Database Name:	labelHeight
Database Class:	Height

Specifies the -height option for the header labels.  This option is only supported by the Tablelist package, but not by Tablelist_tile.

Command-Line Name:	-labelpady
Database Name:	labelPadY
Database Class:	Pad

In the Tablelist package this option specifies the -pady configuration option for the header labels.  In the Tablelist_tile package the value of the -labelpady option is mapped to the corresponding components of the value of the -padding configuration option of the header labels, and the -labelpady option has a theme-specific default value.

Command-Line Name:	-labelrelief
Database Name:	labelRelief
Database Class:	Relief

Specifies the -relief option for the header labels.  This option is different from the standard -relief option defined for the tablelist widget itself.  The default value is raised.

Command-Line Name:	-listvariable
Database Name:	listVariable
Database Class:	Variable

Specifies the name of a variable.  The value of the variable is a list to be displayed inside the widget; if the variable value changes then the widget will automatically update itself to reflect the new value.  The value of the variable must be a valid list.  Each list element corresponds to a row within the widget, and must be a valid list itself; its elements correspond to the cells within the respective row.  Attempts to assign a variable whose value does not fulfil these conditions to -listvariable will cause an error.  Attempts to unset a variable in use as a -listvariable will fail but will not generate an error.

REMARK 1:  For increased efficiency, updating the widget to reflect a changed value of the variable specified with this option is, whenever possible, done at idle time (i.e., when there are no events to process).  On the other hand, most tablelist subcommands make it necessary to perform an immediate update of the widget's internal list according to the value of this variable, before executing the subcommand in question.  Doing this repeatedly can become quite inefficient.  To avoid performance problems, you should always try to separate the operations that build up the value of the variable specified by this option from other commands.  For example, instead of

tablelist::tablelist .tbl ... -listvariable var
set var {}
for {set row 0} {$row < 1000} {incr row} {
    lappend var ...
    .tbl cellconfigure $row,3 -image ...
}

you should write

tablelist::tablelist .tbl ... -listvariable var
set var {}
for {set row 0} {$row < 1000} {incr row} {
    lappend var ...
}
for {set row 0} {$row < 1000} {incr row} {
    .tbl cellconfigure $row,3 -image ...
}

The first method above is quite inefficient, because it requires 1000 updates of the widget's internal list.  The second method performs incomparably faster, because it needs only one synchronization (at the beginning of the second loop).

REMARK 2:  It is not recommended to set this option for a tablelist used as a tree widget, because adding new items to the list specified as its value will result in inserting those list elements into the widget without respecting the tree's internal structure.  There is no problem if you access the variable for reading only, but for that purpose it is more efficient to use the itemlistvar subcommand rather than the -listvariable option.

Command-Line Name:	-movablecolumns
Database Name:	movableColumns
Database Class:	MovableColumns

Specifies a boolean value that determines whether the columns can be moved interactively.  See the DEFAULT AND INDIVIDUAL BINDINGS FOR THE HEADER LABELS section below for information on moving a column interactively.  The default value is 0.

Command-Line Name:	-movablerows
Database Name:	movableRows
Database Class:	MovableRows

Specifies a boolean value that determines whether the rows can be moved interactively.  See the DEFAULT AND INDIVIDUAL BINDINGS FOR THE TABLELIST BODY section below for information on moving a row interactively.  The default value is 0.

Command-Line Name:	-movecolumncursor
Database Name:	moveColumnCursor
Database Class:	MoveColumnCursor

Specifies the mouse cursor to be used when moving a column interactively.  The default value is icon on the windowing systems x11 and win32, and the native cursor closedhand on the Macintosh windowing systems classic and aqua.

Command-Line Name:	-movecursor
Database Name:	moveCursor
Database Class:	MoveCursor

Specifies the mouse cursor to be used when moving a row interactively.  The default value is hand2 on the windowing systems x11 and win32, and the native cursor pointinghand on the Macintosh windowing systems classic and aqua.

Command-Line Name:	-populatecommand
Database Name:	populateCommand
Database Class:	PopulateCommand

Specifies a Tcl command to be invoked by the searchcolumn subcommand before examining the children (or descendants, when used with the -descend option) of a tablelist row whose children have not been inserted yet.  The specified command is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope.  It is expected that this script will insert the children of the row in question, without expanding the node or changing its appearance in any other way.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this option is not supported for Tk versions earlier than 8.3.

REMARK:  There are many similarities between this option and -expandcommand.  Both options are used in the first place to insert children on demand.  The main differences between them are as follows:

1. The command specified by the -populatecommand option is only invoked for rows whose children have not been inserted yet, while the one specified by -expandcommand is always invoked before expanding a row, regardless of whether the children of that row are already present in the widget or not.
2. The command specified by the -expandcommand option may perform visual changes on the node in question, while the role of the one specified by -populatecommand is restricted to inserting the children, without altering the node's appearance in any way.

A logical consequence of the above is that the value of -populatecommand is usually just a stripped-down version of the command specified by the -expandcommand option.  For example, the Directory Viewer demo script might use a command implemented as follows:

proc populateCmd {tbl row} {
    # Get the name of the directory whose leaf name is
    # displayed in the first cell of the specified row
    set dir [$tbl rowattrib $row pathName]

    # Display the contents of the directory $dir
    # as child items of the one identified by $row
    putContents $dir $tbl $row
}

Moreover, the -expandcommand option can be set to a command that invokes the one specified by the -populatecommand option:

proc expandCmd {tbl row} {
    if {[$tbl childcount $row] == 0} {
        populateCmd $tbl $row
    }

    # The rest is just eye candy:
    if {[$tbl childcount $row] != 0} {
        # Update the image displayed in the row's first cell
        $tbl cellconfigure $row,0 -image openFolderImg
    }
}

Command-Line Name:	-protecttitlecolumns
Database Name:	protectTitleColumns
Database Class:	ProtectTitleColumns

Specifies a boolean value that determines whether the boundary of the title column area shall be protected from being crossed when moving a column interactively.  See the DEFAULT AND INDIVIDUAL BINDINGS FOR THE HEADER LABELS section below for information on moving a column interactively.  The default value is 0, specifying that non-title columns can be moved into the title column area and vice-versa.

Command-Line Name:	-resizablecolumns
Database Name:	resizableColumns
Database Class:	ResizableColumns

Specifies a boolean value that determines whether the columns can be resized interactively.  See the DEFAULT AND INDIVIDUAL BINDINGS FOR THE HEADER LABELS section below for information on interactive column resizing.  The default value is 1.

Command-Line Name:	-resizecursor
Database Name:	resizeCursor
Database Class:	ResizeCursor

Specifies the mouse cursor to be used during interactive column resizing.  The default value is sb_h_double_arrow on the windowing systems x11 and win32, and the native cursor resizeleftright on the Macintosh windowing systems classic and aqua.

Command-Line Name:	-selectmode
Database Name:	selectMode
Database Class:	SelectMode

Specifies one of several styles for manipulating the selection.  The value of the option may be arbitrary, but the default bindings expect it to be either single, browse, multiple, or extended.  The default value is browse.

Command-Line Name:	-selecttype
Database Name:	selectType
Database Class:	SelectType

Specifies one of two selection types for the tablelist widget: row or cell.  If the selection type is row then the default bindings will select and deselect entire items, and the whole row having the location cursor will be displayed as active when the tablelist has the keyboard focus.  If the selection type is cell then the default bindings will select and deselect individual elements, and the single cell having the location cursor will be displayed as active when the tablelist has the keyboard focus.  The default value is row.

Command-Line Name:	-setfocus
Database Name:	setFocus
Database Class:	SetFocus

Specifies a boolean value that determines whether a click with the left mouse button anywhere into the tablelist's body, including the separators and the embedded images (more precisely, any descendants of the tablelist widget having the binding tag TablelistBody) should set the focus to the body of the tablelist widget if the latter's state is normal.  The default value is 1.

Command-Line Name:	-showarrow
Database Name:	showArrow
Database Class:	ShowArrow

Specifies a boolean value that determines whether the sortbycolumn and sortbycolumnlist subcommands of the Tcl command associated with the widget should place an arrow indicating the sort order into the header label(s) of the column(s) specified by their first argument.  The default value is 1.

Command-Line Name:	-showeditcursor
Database Name:	showEditCursor
Database Class:	ShowEditCursor

Specifies a boolean value indicating whether the mouse cursor should change automatically to one having the shape of a pencil whenever a click with mouse button 1 would start an editing session.  The default value is 1.

REMARK:  On X11 and Mac OS X Aqua, the cursor set by Tablelist when a mouse klick would start the interactive cell editing, is Tk's built-in cursor pencil.  Since on Windows this cursor is ugly and not really usable, the edit cursor shown by Tablelist on this platform will be the one loaded from the file pencil.cur, located in the scripts subdirectory of the Tablelist installation directory.  If the application was started as a starpack containing the Tablelist package, then the edit cursor will be loaded from a copy of this file, created automatically by Tablelist in the user's Temp directory.

The file pencil.cur is a copy of the file  Ubuntu Handwriting.cur,  downloaded from the address

http://www.rw-designer.com/cursor-detail/46745

Command-Line Name:	-showhorizseparator
Database Name:	showHorizSeparator
Database Class:	ShowHorizSeparator

Specifies a boolean value that controls whether to place a horizontal separator just below the last row if the value of the -showseparators option is true, that of the -fullseparators option is false, and there is free space left between the last row and the bottom of the tablelist window.  The default value is 1.

Command-Line Name:	-showlabels
Database Name:	showLabels
Database Class:	ShowLabels

Specifies a boolean value that determines whether the header labels are to be shown or not.  The default value is 1.

Command-Line Name:	-showseparators
Database Name:	showSeparators
Database Class:	ShowSeparators

Specifies a boolean value that determines whether the columns are to be separated with borders.  The default value is 0.  The separators are implemented as thin frames with sunken relief in the package Tablelist, and as tile separator widgets in the package Tablelist_tile.  They are attached to the right edges of the header labels, and are only created if the value of this option is true.  In this case, a horizontal separator will also be created and placed just below the last row, provided that the value of the -showhorizseparator option is true, that of the -fullseparators option is false, and there is free space left between the last row and the bottom of the tablelist window.

REMARK:  Tablelist doesn't support horizontal separators (except for the one mentioned above), but a nice distinguishing effect for the rows can be achieved with the aid of the -stripebackground option discussed below.

Command-Line Name:	-snipstring
Database Name:	snipString
Database Class:	SnipString

Specifies the string to be used as snip indicator when displaying the elements that don't fit into their cells.  The default is an ellipsis ("...").

Command-Line Name:	-sortcommand
Database Name:	sortCommand
Database Class:	SortCommand

Specifies a command to be used for the comparison of the items when invoking the sort subcommand of the Tcl command associated with the tablelist widget.  To compare two items (viewed as lists of cell contents within one row each) during the sort operation, the command is automatically concatenated with the two items and the resulting script is evaluated.  The script should return an integer less than, equal to, or greater than zero if the first item is to be considered less than, equal to, or greater than the second, respectively.

Command-Line Name:	-spacing
Database Name:	spacing
Database Class:	Spacing

Specifies additional space to provide above and below each row of the widget.  The option's value may have any of the standard forms for screen distances.  It defaults to 0.  See also the -tight option.

Command-Line Name:	-state
Database Name:	state
Database Class:	State

Specifies one of two states for the tablelist widget: normal or disabled.  If the widget is disabled then neither items nor columns may be inserted, deleted, updated, or moved, the items, header labels, and the up- or down-arrow are drawn in the -disabledforeground, -labeldisabledforeground, and -arrowdisabledcolor color, respectively, the selection cannot be modified and is not shown (although the selection information is retained), the header labels are completely insensitive, and no interactive cell editing can be performed.  In addition, in disabled state any color options specified at column, row, or cell level will be ignored.

Command-Line Name:	-stretch
Database Name:	stretch
Database Class:	Stretch

Specifies the columns to be stretched in order to fill the tablelist window if necessary.  The option's value may be all or a list of column indices in any of the forms described in the COLUMN INDICES section below.  In the second case, the specified column indices are replaced with their numerical equivalents, except for end and last, which are viewed as dynamic column indices whose numerical equivalent (i.e., the index of the tablelist's last column) might change during program execution and therefore will be recomputed every time the columns are stretched.  The list will be updated automatically whenever columns are inserted, deleted, or moved.  The number of pixels by which a column is stretched is proportional to its width in pixels.  The default value of this option is an empty list, meaning that no column will be stretched to eliminate the blank space that might appear at the right of the table.  (Note that the blank space following the header labels is filled with a dummy, insensitive label having the same background, borderwidth, and relief as the "normal" header labels.)  This option is ignored if the value of the -width configuration option is zero or less.

Command-Line Name:	-stripebackground or -stripebg
Database Name:	stripeBackground
Database Class:	Background

Specifies the background color to use when displaying the items belonging to a stripe.  Each stripe is composed of the same number stripeHeight of consecutive viewable items, according to the value of the -stripeheight configuration option.  The first stripeHeight viewable items are "normal" ones; they are followed by a stripe composed of the next stripeHeight viewable items, which in turn is followed by the same number of "normal" viewable items, and so on.  In the Tablelist package and in most themes supported by Tablelist_tile, the default value is an empty string, indicating that the stripes will inherit the background color specified by the -background configuration option.  When using Tablelist_tile with the tileqt theme then the default value is given by the global KDE option alternateBackground, which in turn depends on the current color scheme.  In this case it is recommended to either keep that default value retrieved from KDE, or to use an explicitly specified empty string if no stripes are to be displayed.  The -stripebackground option has a higher priority than the -background column configuration option, but a lower priority than the -background option specified at row or cell level.

Command-Line Name:	-stripeforeground or -stripefg
Database Name:	stripeForeground
Database Class:	Foreground

Specifies the foreground color to use when displaying the items belonging to a stripe.  Each stripe is composed of the same number stripeHeight of consecutive viewable items, according to the value of the -stripeheight configuration option.  The first stripeHeight viewable items are "normal" ones; they are followed by a stripe composed of the next stripeHeight viewable items, which in turn is followed by the same number of "normal" viewable items, and so on.  The default value is an empty string, indicating that the stripes will inherit the foreground color specified by the -foreground configuration option.  The -stripeforeground option has a higher priority than the -foreground column configuration option, but a lower priority than the -foreground option specified at row or cell level.

Command-Line Name:	-stripeheight
Database Name:	stripeHeight
Database Class:	StripeHeight

Specifies the number of items in each stripe.  If zero or less then no stripes are displayed.  The default is 1.

Command-Line Name:	-takefocus
Database Name:	takeFocus
Database Class:	TakeFocus

This option determines whether the widget accepts the focus during keyboard traversal.  It is almost identical to the standard option of the same name (see the options manual entry for details).  The only difference is that not the widget itself but its body child (containing the items) will receive the focus during keyboard traversal with the standard keys (Tab and Shift-Tab).

Command-Line Name:	-targetcolor
Database Name:	targetColor
Database Class:	TargetColor

Specifies the color of the gap displayed in the tablelist's body or header to indicate the target position when moving a row or column interactively.  The target indicator belonging to the tablelist's body can also be displayed explicitly with the aid of the showtargetmark subcommand, and its path name can be retrieved via targetmarkpath.  The default value is black.

Command-Line Name:	-tight
Database Name:	tight
Database Class:	Tight

Specifies a boolean value that determines whether to eliminate the one-pixel additional space left below each tablelist row for Tk listbox compatibility.  The default value is 0, which draws a one-pixel additional space below each row, just like a Tk core listbox.

REMARK:  You can set this option to true and at the same time provide additional space above and below each row with the aid of the -spacing configuratiom option.  As a result, the space below each row will be the same as above the row (and not one pixel higher).

Command-Line Name:	-titlecolumns
Database Name:	titleColumns
Database Class:	TitleColumns

Specifies the number of the non-scrollable columns at the left edge of the window, also called title columns.  The positions of these columns will not change when adjusting the horizontal view by invoking the scan, seecell, seecolumn, or xview subcommand.  The default value is 0.  The value of this option also determines the scrolling unit used by the commands mentioned above when shifting the horizontal view: if it is positive then the horizontal scrolling is performed column-wise, otherwise by character units (the width of the 0 character).

The end of the title column area is visualized with the aid of a separator, attached to the right edge of the header label corresponding to the last non-hidden title column.  This special separator is always displayed to mark the end of the title columns (if any), independently of the value of the -showseparators option.  The user can easily distinguish it from the other separators by means of its background color, which is different from that of the other separators.

For technical reasons (the use of the -elide option for a text widget tag), this option is not supported for Tk versions earlier than 8.3.

Command-Line Name:	-tooltipaddcommand
Database Name:	tooltipAddCommand
Database Class:	TooltipAddCommand

Specifies a Tcl command to be used for displaying cell- and column label-specific balloon help.  When the mouse pointer enters a cell, the command is automatically concatenated with the name of the tablelist widget and the cell's row and column indices, and the resulting script is evaluated in the global scope.  Similarly, when the mouse pointer enters a header label, the command is automatically concatenated with the name of the tablelist widget, the number -1, and the column index of the respective label, and the resulting script is evaluated in the global scope.  In both cases, the action described above is only triggered if both the value of this option and that of -tooltipdelcommand are nonempty strings.

For example, consider the procedure tooltipAddCmd shown below, which makes use of the DynamicHelp::add command from the BWidget package to display the full cell and label texts as tooltips for the cells and header labels with snipped contents.

proc tooltipAddCmd {tbl row col} {
    if {($row >= 0 && [$tbl iselemsnipped $row,$col fullText]) ||
        ($row <  0 && [$tbl istitlesnipped $col fullText])} {
        DynamicHelp::add $tbl -text $fullText
    }
}

A tablelist widget can use this procedure by specifying

... -tooltipaddcommand tooltipAddCmd -tooltipdelcommand DynamicHelp::delete

If you prefer to use the tooltip::tooltip command from the tooltip package contained in tklib then the procedure becomes

proc tooltipAddCmd {tbl row col} {
    if {($row >= 0 && [$tbl iselemsnipped $row,$col fullText]) ||
        ($row <  0 && [$tbl istitlesnipped $col fullText])} {
        tooltip::tooltip $tbl $fullText
    }
}

and can be used by specifying

... -tooltipaddcommand tooltipAddCmd -tooltipdelcommand "tooltip::tooltip clear"

Please note that in the less common case that the name of your tablelist widget contains spaces, the  tooltip::tooltip clear  command won't work as expected.  As a workaround you can use the slightly modified approach shown below:

proc tooltipDelCmd tbl { tooltip::tooltip $tbl "" }

... -tooltipaddcommand tooltipAddCmd -tooltipdelcommand tooltipDelCmd

Both examples above make use of the iselemsnipped and istitlesnipped subcommands, to make sure that the full cell and label texts will only be displayed for those cells and header labels whose contents are snipped.

Command-Line Name:	-tooltipdelcommand
Database Name:	tooltipDelCommand
Database Class:	TooltipDelCommand

Specifies a Tcl command to be used for removing the cell- or column label-specific balloon help.  When the mouse pointer leaves a cell or a header label, the command specified by this option is automatically concatenated with the name of the tablelist widget and the resulting script is evaluated in the global scope.  This action is only triggered if both the value of this option and that of -tooltipaddcommand are nonempty strings.  Common values for this option are "DynamicHelp::delete" (which requires the BWidget package) and  "tooltip::tooltip clear"  (which requires the tooltip package contained in tklib).  Their usage is shown in the examples above.

Command-Line Name:	-treecolumn
Database Name:	treeColumn
Database Class:	TreeColumn

Specifies the column to contain the indentations and expand/collapse controls for a tablelist used as a tree widget.  The option's value may be a column index in any of the forms described in the COLUMN INDICES section below.  The specified column index is replaced with its numerical equivalent, and it will be updated automatically whenever columns are inserted, deleted, or moved.  The default value is 0.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this option is not supported for Tk versions earlier than 8.3.

REMARK:  Please note that the tree structure will only be displayed as expected if the column specified by this option is left-aligned.  It is your responsibility to make sure that this restriction is fulfilled when using a tablelist as a tree widget.

Command-Line Name:	-treestyle
Database Name:	treeStyle
Database Class:	TreeStyle

Specifies the look & feel of the column containing the indentations and expand/collapse controls for a tablelist used as a tree widget.  This includes, among others, the images used for displaying the expand/collapse controls, the indentation width, and whether expand/collapse controls and indentations are to be protected when selecting a row or cell.  The currently supported values are shown at a glance below:

aqua	gtk	adwaita	ubuntu
mint	mate	newWave
plastique	oxygen1	oxygen2	klearlooks
plastik	baghira	phase
winnative	winxpBlue	winxpOlive	winxpSilver
vistaAero	vistaClassic	win7Aero	win7Classic
ambiance	dust	dustSand	radiance

If the tree style is aqua, gtk, adwaita, ubuntu, mint, mate, newWave, oxygen2, vistaAero, or win7Aero, and the Tk version is either 8.6 (with built-in PNG support) or 8.5 and the img::png package can be loaded into the interpreter, then the images used for displaying the expand/collapse controls are PNG images with alpha channel.  Otherwise (i.e., for the other tree styles or in the absence of PNG support) GIF images are used for the expand/collapse controls.

The following table contains a detailed description of the tree styles, in alphabetical order:

Value	Screenshot	Comments
adwaita		Inspired by the GTK+ 3 theme Adwaita.
ambiance		Inspired by the GTK+ theme Ambiance.  Recommended to be used with large fonts and images.
aqua		Tablelist:  Default for the windowing systems aqua and classic. Tablelist_tile:  Default for the themes aqua and Aquativo.
baghira		Tablelist_tile:  Default for the Qt styles baghira, cde, and motif within the tileqt theme.  Also used by some flavors of qtcurve.
dust		Inspired by the GTK+ theme Dust.  Recommended to be used with large fonts and images.
dustSand		Inspired by the GTK+ theme Dust Sand.  Recommended to be used with large fonts and images.
gtk		Tablelist:  Default for the windowing system x11. Tablelist_tile:  Default for the themes blue, clam, classic, default, kroc, sriv, srivlg, and step, as well as for the Qt style gtk+ within the tileqt theme.
klearlooks		Inspired by the KDE 4 style Klearlooks. Tablelist_tile:  Default for the Qt style qtcurve within the tileqt theme.
mate		Inspired by a few GTK+ themes of the MATE desktop.
mint		Inspired by a few GTK+ 3 themes bundled with the Linux Mint distribution.
newWave		Inspired by the GTK+ theme New Wave.
oxygen1		Inspired by the KDE 4 style Oxygen.
oxygen2		Tablelist_tile:  Default for the Qt style oxygen within the tileqt theme.
phase		Tablelist_tile:  Default for the Qt style phase within the tileqt theme.
plastik		Tablelist_tile:  Default for the plastik theme and the Qt style plastik within the tileqt theme.
plastique		Tablelist_tile:  Default for the Qt style plastique within the tileqt theme.
radiance		Inspired by the GTK+ theme Radiance.  Recommended to be used with large fonts and images.
ubuntu		Inspired by a couple of GTK+ 3 themes bundled with the Ubuntu Linux distribution.
vistaAero		Tablelist:  Default for the windowing system win32 on Windows Vista. Tablelist_tile:  Default for the vista theme on Windows Vista with the Vista Aero style.
vistaClassic		Tablelist:  Default for the windowing system win32 on Windows Vista with the Windows Classic style. Tablelist_tile:  Default for the vista theme on Windows Vista with the Windows Classic style.
win7Aero		Tablelist:  Default for the windowing system win32 on Windows 7. Tablelist_tile:  Default for the vista theme on Windows 7 with the Windows 7 Aero style.
win7Classic		Tablelist:  Default for the windowing system win32 on Windows 7 with the Windows Classic style. Tablelist_tile:  Default for the vista theme on Windows 7 with the Windows Classic style.
winnative		Tablelist:  Default for the windowing system win32 on Windows 2000 and Windows XP with the Windows Classic style. Tablelist_tile:  Default for the themes alt, keramik, keramik_alt, winnative, and xpnative with the Windows Classic style, as well as for most Qt styles within the tileqt theme.
winxpBlue		Tablelist:  Default for the windowing system win32 on Windows XP with the Blue color scheme. Tablelist_tile:  Default for the winxpblue theme and the xpnative theme on Windows XP with the Blue color scheme.
winxpOlive		Tablelist:  Default for the windowing system win32 on Windows XP with the Olive Green color scheme. Tablelist_tile:  Default for the xpnative theme on Windows XP with the Olive Green color scheme.
winxpSilver		Tablelist:  Default for the windowing system win32 on Windows XP with the Silver color scheme. Tablelist_tile:  Default for the xpnative theme on Windows XP with the Silver color scheme.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this option is not supported for Tk versions earlier than 8.3.

Command-Line Name:	-width
Database Name:	width
Database Class:	Width

Specifies the desired width for the window, in average-size characters of the widget's font.  If zero or less then the desired width for the window is made just large enough to hold all the columns in the tablelist widget.

COLUMN CONFIGURATION OPTIONS

The following options are currently supported by the columncget, columnconfigure, configcolumnlist, and configcolumns commands:

-align alignment

Specifies how to align the elements of the column.  It must be one of left, right, or center.  This option also refers to the column's title if the -labelalign option hasn't been specified for the given column, or if its value is an empty string.  The -align option is tied to the alignment element corresponding to this column in the list specifying the value of the -columns option for the tablelist widget; changes in either will automatically be reflected in the other.

-background color or -bg color

Specifies the normal background color to use when displaying the contents of the column.

-changesnipside boolean

Specifies whether to override the alignment-specific default position of the snip indicator when displaying the elements of the column (excluding its title).  The default value is 0, meaning that the snip string will be appended to the elements if the column's alignment is left or center and prepended to them in case the alignment is right.

-changetitlesnipside boolean

Specifies whether to override the alignment-specific default position of the snip indicator when displaying the column's title.  The default value is 0, meaning that the snip string will be appended to the title if the column's alignment is left or center and prepended to it in case the alignment is right.

-editable boolean

Specifies whether the elements of the column can be edited interactively.  The default value is 0.  The value of this option can be overridden for individual cells by using the cell configuration option of the same name.

-editwindow name

Specifies the type of the temporary embedded widget to be used for interactive editing of the contents of the given column's cells.  name may be one of entry (which is the default), text, spinbox (the latter for Tk versions 8.4 or higher), checkbutton, menubutton, ttk::entry, ttk::spinbox, ttk::combobox, ttk::checkbutton, or ttk::menubutton (the latter five only in the presence of the tile widget engine), or the value returned by one of the registration commands for widgets from the packages BWidget, Iwidgets, combobox (by Bryan Oakley), ctext, and Mentry (or Mentry_tile).  For example, you can use  -editwindow ComboBox  after registering the ComboBox widget for interactive cell editing with the aid of the tablelist::addBWidgetComboBox command.  Similarly, you can use  -editwindow combobox  after registering Bryan Oakley's combobox widget for interactive cell editing by invoking the tablelist::addOakleyCombobox command.  The value of this option can be overridden for individual cells by using the cell configuration option of the same name.

-font font

Specifies the font to use when displaying the contents of the column.

-foreground color or -fg color

Specifies the normal foreground color to use when displaying the contents of the column.

-formatcommand command

Specifies a Tcl command to be invoked when displaying the content of a cell within this column or adding it to the selection when the latter is being exported.  If command is a nonempty string, then it is automatically concatenated with the cell's text, the resulting script is evaluated in the global scope, and the return value is displayed in the cell or added to the selection instead of the original data.

For example, if a time value in seconds is being inserted into the cell and command is the procedure formatDate defined as

proc formatDate clockVal {
    return [clock format $clockVal -format "%Y-%m-%d"]
}

then the text displayed in the cell will be the date in the specified format, not the time value in seconds.

This option is also used by Tablelist when building the return values of the getformatted, getformattedcolumns, and getformattedcells subcommands, or searching for a text pattern passed to the searchcolumn subcommand with the -formatted option.  All the other subcommands, notably get, getcolumns, getcells, rowcget, columncget, cellcget, sort, sortbycolumn, sortbycolumnlist, and refreshsorting operate on the original cell text, which is contained in the widget's internal list.  In the case of the above example, this will make it possible to sort the items quite easily by time, with a second's precision, even if their visual representation only contains the year, month, and day.

The -formatcommand option comes in handy if only images or embedded windows are to be displayed in a column but the texts associated with the cells may not simply be empty strings because they are needed for other purposes (like sorting or editing).  In such cases, a procedure returning an empty string can be used as the option's value, thus making sure that the textual information contained in that column remains hidden.

The demo scripts included in the Tablelist distribution contain further examples demonstrating the use of this option.  The most interesting ones are also described in Tablelist Programmer's Guide.

In the more sophisticated case that the result of the formatting should also depend on the cell's row, you will have to invoke the formatinfo subcommand, which provides the necessary information about the cell whose content is being formatted.

Due to the -formatcommand column configuration option, we will distinguish between the internal value of a tablelist element and its formatted version.  The latter is the result of the invocation of the script corresponding to the -formatcommand option of the element's column, or the element itself if this option was not set for that column.  Consequently, the formatted version of a tablelist item is a list comprised of the formatted elements of the original item.

-hide boolean

Specifies whether to hide the column when displaying the widget or exporting its selection.  The default value is 0.  After toggling the hidden state of a column, the <<TablelistColHiddenStateChanged>> virtual event is generated, with its -data option set to the numerical column index for Tk versions 8.5 or higher.

-labelalign alignment

Specifies how to align the column's title.  It must be one of left, right, or center, or an empty string.  If this option hasn't been specified for the given column, or if its value is an empty string, then the header title will have the same alignment as the elements of the column, as given by the -align column configuration option or by the alignment element corresponding to this column in the list specifying the value of the -columns global option.

-labelbackground color or -labelbg color
-labelborderwidth screenDistance or -labelbd screenDistance
-labelcommand command
-labelcommand2 command
-labelfont fontName
-labelforeground color or -labelfg color
-labelheight lines
-labelpady screenDistance
-labelrelief relief

The value of each of these options may also be an empty string.  These options are the column-specific equivalents of the global ones having the same names, described in the WIDGET-SPECIFIC OPTIONS section.  They override the options of the same names set at widget level if the specified value is not empty.  If one of these options hasn't been specified for the given column, or if its value is an empty string, then that option will not be used at column level; the global option of the same name will be used instead.  The -labelactivebackground, -labelactiveforeground, and -labeldisabledforeground options are only defined at widget level; there are no column configuration options with these names.  The -labelbackground and -labelheight options are only supported by the Tablelist package, but not by Tablelist_tile.

-labelimage image

Specifies the name of the Tk image to be displayed in the header label.  image must be the result of an invocation of the  image create  command, or an empty string, specifying that no image is to be displayed.  If the label's text is right-aligned then the image will be displayed to the right of the text, otherwise to its left.  The text and the image are separated from each other by a gap corresponding to the width of a space character in the given label's font.

-maxwidth width

width must be a number.  A positive value specifies the column's maximum width in average-size characters of the widget's font.  If width is negative, its absolute value is interpreted as a maximum column width in pixels.  Finally, a value of zero (which is the default) specifies that the column's maximum width is to be made just large enough to hold all the elements in the column, including its header.  This option is only relevant if the given column has dynamic width, i.e., if its width was set to 0.

-name name

Specifies a name associated with the column.  While the numerical index of a column might change by inserting, deleting, or moving columns, its name remains constant and can be used as a safe alternative column index (see the COLUMN INDICES section for details).  Similarly, it can also be used as the second component of a cell index of the form row,col, as described in the CELL INDICES section.  To avoid ambiguities, column names should be different from any other forms of column indices (like numbers, active, anchor, end, last, left, right, or any of their abbreviations).  They should also be different from (any abbreviations of) the string all, which may be specified as the value of the -stretch configuration option.  The default value is an empty string.

-resizable boolean

Specifies whether the column can be resized interactively.  See the DEFAULT AND INDIVIDUAL BINDINGS FOR THE HEADER LABELS section for information on interactive column resizing.  The default value is 1.  This option is only relevant if the value of the -resizablecolumns widget configuration option is true.

-selectbackground color

Specifies the background color to use when displaying the contents of a cell in the given column while the cell is selected.

-selectforeground color

Specifies the foreground color to use when displaying the contents of a cell in the given column while the cell is selected.

-showarrow boolean

Specifies whether the sortbycolumn command with the given column index as first argument and the sortbycolumnlist command having the given column index as element of its first argument should place an arrow indicating the sort order into the column's label.  The default value is 1.  This option is only relevant if the value of the -showarrow widget configuration option is true.

-showlinenumbers boolean

Specifies whether the given column should display the line numbers (starting with 1 and ending with the number of the non-hidden rows).  The default value is 0.

The following details assume that the given column's -showlinenumbers option was set to true:  Associating the line numbers with the non-hidden rows takes place automatically whenever items are inserted, deleted, updated, moved, or sorted, or their -hide option is toggled.  For increased efficiency, this is done at idle time.  For example, if several items are inserted into or deleted from the tablelist widget, then the necessary renumbering of the non-hidden rows will be performed as an idle callback, the next time the event loop is entered and there are no events to process.  The line numbers will override any previous contents of the column's cells.  They are, per default, displayed without leading zeros, but this (and the display format in general) can be changed with the aid of the -formatcommand column configuration option.

The sortbycolumn and sortbycolumnlist subcommands as well as the tablelist::sortByColumn and tablelist::addToSortColumns commands check the column indices passed to them as arguments and don't perform any sorting by those columns that have been configured to display the line numbers (see the corresponding descriptions for details).

REMARK:  Please note that the line numbers refer to the non-hidden rows, which in case of a tablelist used as a tree widget might be different from the viewable ones.  Consequently, collapsing a row will, in general, not lead to renumbering the lines (instead, it will just elide the corresponding range of line numbers).

-sortcommand command

This option is only relevant if the value of the -sortmode option for the given column is command.  It specifies a command to be used for the comparison of the column's elements when invoking the sortbycolumn command with the given column index as first argument or the sortbycolumnlist command having the given column index as element of its first argument.  To compare two elements during the sortbycolumn or sortbycolumnlist operation, command is automatically concatenated with the two elements and the resulting script is evaluated.  The script should return an integer less than, equal to, or greater than zero if the first element is to be considered less than, equal to, or greater than the second, respectively.

-sortmode mode

Specifies how to compare the column's elements when invoking the sortbycolumn command with the given column index as first argument or the sortbycolumnlist command having the given column index as element of its first argument.  mode may have any of the following values:

ascii	Use string comparison with Unicode code-point collation order (the name is for backward-compatibility reasons).  This is the default.
asciinocase	This is the same as ascii, except that comparisons are handled in a case-insensitive manner.
command	Use the command specified by the -sortcommand column configuration option to compare the column's elements.
dictionary	Use dictionary-style comparison.  This is the same as ascii, except: (a) case is ignored except as a tie-breaker; (b) if two strings contain embedded numbers, the numbers compare as integers, not characters.  For example, bigBoy sorts between bigbang and bigboy, and x10y sorts between x9y and x11y.
integer	Convert the elements to integers and use integer comparison.
real	Convert the elements to floating-point values and use floating-point comparison.

-stretchable boolean

Specifies whether the column is to be stretched in order to fill the tablelist window if necessary.  The value of this option is tied to that of the -stretch option for the tablelist widget; changes in either will automatically be reflected in the other.

-stripebackground color

Specifies the background color to use when displaying the contents of a cell in the given column if the cell's row belongs to a stripe.

-stripeforeground color

Specifies the foreground color to use when displaying the contents of a cell in the given column if the cell's row belongs to a stripe.

-text list

Specifies a list of strings to be displayed in the cells of the given column, i.e., updates the elements contained in the column.  If the tablelist's state is disabled then this option will be ignored.

-title title

Specifies the text to be displayed in the column's header.  This option is tied to the title element corresponding to the given column in the list specifying the value of the -columns option for the tablelist widget; changes in either will automatically be reflected in the other.

-valign verticalAlignment

Specifies the vertical alignment of embedded images and windows displayed in the given column.  The verticalAlignment must be one of center (which is the default), top, or bottom.  Especially in rows containing multi-line elements, embedded images or windows are often not as tall as the rows themselves; in such cases, this option enables you to control the column-dependent vertical alignment of these embedded objects.  The value of this option can be overridden for individual cells by using the cell configuration option of the same name.

-width width

width must be a number.  A positive value specifies the column's width in average-size characters of the widget's font.  If width is negative, its absolute value is interpreted as a column width in pixels.  Finally, a value of zero specifies that the column's width is to be made just large enough to hold all the elements in the column, including its header (but no larger than the maximum width indicated by the -maxwidth column configuration option).  This option is tied to the width element corresponding to the given column in the list specifying the value of the -columns option for the tablelist widget; changes in either will automatically be reflected in the other.

-wrap boolean

Specifies whether to display those elements of the given column that don't fit into their cells in word-wrapped multi-line rather than snipped form.  The default value is 0.  If the specified column has static width or a nonzero -maxwidth value and the value of this option is true, then elements of the column that are too long to be displayed in a single line will be broken up into several lines.  The same applies to the individual lines of the multi-line elements (i.e., elements containing newline characters): they will also be wrapped if necessary, thus giving rise to additional line breaks.  In both cases, the line breaks are chosen at word boundaries wherever possible, and they are only used for the external representation of the strings contained in the given column, without affecting the internal contents of the cells.

The -background, -font, -foreground, -selectbackground, -selectforeground, -stripebackground, and -stripeforeground column configuration options override the options of the same names set at widget level (but not the ones set at cell or row level) if the specified value is not an empty string.  See the COLORS AND FONTS section for further details on these options.

ROW CONFIGURATION OPTIONS

The following options are currently supported by the rowcget, rowconfigure, configrowlist, and configrows commands:

-background color or -bg color

Specifies the normal background color to use when displaying the contents of the row.

-font font

Specifies the font to use when displaying the contents of the row.

-foreground color or -fg color

Specifies the normal foreground color to use when displaying the contents of the row.

-hide boolean

Specifies whether to hide the row when displaying the widget or exporting its selection.  The default value is 0.  After toggling the hidden state of a row, the <<TablelistRowHiddenStateChanged>> virtual event is generated, with its -data option set to the numerical row index for Tk versions 8.5 or higher.

For technical reasons (the use of the -elide option for a text widget tag), this option is not supported for Tk versions earlier than 8.3.

CAUTION:  Tk versions 8.3 - 8.4.12 had a bug that caused a segmentation fault if the whole content of a text widget was elided.  This bug was also present in Tk 8.5.a1 - 8.5.a3.  When using one of these earlier Tk versions, this bug will produce a crash if all the rows of a tablelist widget are hidden.  It is your responsibility to avoid such situations when using a Tk version having this bug!

-name name

Specifies a name associated with the row.  While the numerical index of a row might change by inserting, deleting, or moving rows, or by sorting the items, its name remains constant and can be used as a safe alternative row index (see the ROW INDICES section for details).  Similarly, it can also be used as the first component of a cell index of the form row,col, as described in the CELL INDICES section.  To avoid ambiguities, row names should be different from any other forms of row indices (like numbers, full keys, active, anchor, end, last, top, bottom, or any of their abbreviations).  The default value is an empty string.

-selectable boolean

Specifies whether the elements displayed in the given row can be selected.  The default value is 1.  If the value 0 was given then any attempt to select the item contained in this row with the aid of the  selection set  widget command or any of its elements by using the  cellselection set  command will be silently ignored; moreover, an existing old (cell) selection is removed from the row.

-selectbackground color

Specifies the background color to use when displaying the contents of a cell in the given row while the cell is selected.

-selectforeground color

Specifies the foreground color to use when displaying the contents of a cell in the given row while the cell is selected.

-text list

Specifies a list of strings to be displayed in the cells of the given row, i.e., updates the item contained in the row.  If the tablelist's state is disabled then this option will be ignored.

The -background, -font, -foreground, -selectbackground, and -selectforeground row configuration options override the options of the same names set at column or widget level (but not the ones set at cell level) if the specified value is not an empty string.  See the COLORS AND FONTS section for further details on these options.

CELL CONFIGURATION OPTIONS

The following options are currently supported by the cellcget, cellconfigure, configcelllist, and configcells commands:

-background color or -bg color

Specifies the normal background color to use when displaying the contents of the cell.

-editable boolean

Specifies whether the contents of the cell can be edited interactively.  The default value is 0.  This option overrides the one of the same name for the column containing the given cell.

-editwindow name

Specifies the type of the temporary embedded widget to be used for interactive editing of the cell's contents.  The default value is entry.  This option overrides the one of the same name for the column containing the given cell, and may have the same values as its column-related counterpart.

-font font

Specifies the font to use when displaying the contents of the cell.

-foreground color or -fg color

Specifies the normal foreground color to use when displaying the contents of the cell.

-image image

Specifies the name of the Tk image to be displayed in the cell.  image must be the result of an invocation of the  image create  command, or an empty string, specifying that no image is to be displayed.  If the column containing the cell is right-aligned then the image will be displayed to the right of the cell's text, otherwise to its left.  The text and the image are separated from each other by a gap of 4 pixels.  If for the same cell the -window option was specified with a nonempty value then it overrides the -image option.  If the tablelist's state is disabled then this option will be ignored.

To display an image in a cell, Tablelist makes use of an embedded label widget (which is created on demand).  This requires more memory than inserting the image directly into the tablelist's body, but has the main advantage of making it possible to adjust the width of the label containing the widget to fit into its column.  This has the visual effect of cutting off part of the image from its right side.  To make sure that images with transparent background will be displayed correctly, the background color of the label widgets containing the embedded images is automatically updated whenever necessary.

-selectbackground color

Specifies the background color to use when displaying the contents of the cell while it is selected.

-selectforeground color

Specifies the foreground color to use when displaying the contents of the cell while it is selected.

-stretchwindow boolean

Specifies whether the width of the window embedded into the cell shall be adapted dynamically to the width of the cell's column if the latter is non-zero (i.e., static).  The default value is 0.  If the value of this option is true and the column was specified with a non-zero width or was resized interactively, then the width of the embedded window (if any) will be adjusted automatically so the window fills the whole horizontal space belonging to that column (except the left and right margins).  Please note that in this case the cell's text will remain hidden.  On the other hand, if the column is of dynamic width then this option will be ignored and both the cell's text and its embedded window (if any) will be displayed as usual.  The easiest way to avoid this discrepancy is to set the cell's text to an empty string or make sure that the column's elements are always displayed as empty strings, by using the -formatcommand column configuration option.

-text text

Specifies the string to be displayed in the given cell, i.e., updates the element contained in the cell.  If the tablelist's state is disabled then this option will be ignored.

-valign verticalAlignment

Specifies the vertical alignment of the embedded image or window displayed in the cell (if any).  The default value is center.  This option overrides the one of the same name for the column containing the given cell, and may have the same values as its column-related counterpart.

-window command

Specifies a Tcl command creating the window to be embedded into the cell.  The command is automatically concatenated with the name of the tablelist widget, the cell's row and column indices, as well as the path name of the embedded window to be created, and the resulting script is evaluated in the global scope.  command may also be an empty string, specifying that no embedded window is to be displayed.  If the column containing the cell is right-aligned then the window will be displayed to the right of the cell's text, otherwise to its left.  The text and the window are separated from each other by a gap of 4 pixels.  If this option was specified with a nonempty value then it overrides the -image cell configuration option.  If the tablelist's state is disabled then this option will be ignored.

REMARK:  There are several situations where the embedded window will be destroyed and later recreated by invoking the script mentioned above.  For example, when changing the value of some of the tablelist widget or column configuration options, sorting the items, or moving a row or a column, the widget's contents will be redisplayed, which makes it necessary to recreate the embedded windows.  This operation won't preserve the changes made on the embedded windows after their creation.  For this reason, you should avoid any changes on embedded windows outside their creation scripts.

-windowdestroy command

Specifies a Tcl command to be invoked when a window embedded into the cell is destroyed.  The command is automatically concatenated with the name of the tablelist widget, the cell's row and column indices, as well as the path name of the embedded window, and the resulting script is evaluated in the global scope.  This option provides an easy-to-use alternative to binding a script to the embedded window's <Destroy> event from within its creation script, specified as the value of the -window cell configuration option.

-windowupdate command

Specifies a Tcl command responsible for adapting the background and foreground colors of a window embedded into the cell to the current background and foreground colors of the cell itself.  The current cell colors depend on whether the tablelist widget's state is normal or disabled, whether the cell is selected, and whether its row is displayed with the colors specified by the -stripebackground and -stripeforeground widget or column configuration options.  The command is automatically concatenated with the name of the tablelist widget, the cell's row and column indices, the path name of the embedded window, the option name -background and the cell's current background color, as well as the option name -foreground and the cell's current foreground color.  The resulting script is then evaluated in the global scope.  The specified command can use the colors passed to it as arguments to set the corresponding colors of the embedded window to the same values.

For example, if the embedded window is a text or ctext widget and you want for its background and foreground colors to be automatically adapted to the cell's current background and foreground colors, then you can achieve this by setting the -windowupdate cell configuration option to updateWindow, where the latter is the procedure implemented as follows:

proc updateWindow {tbl row col w args} {
    eval [list $w configure] $args
}

With Tcl/Tk 8.5 or above, you can use the more compact form

proc updateWindow {tbl row col w args} {
    $w configure {*}$args
}

In this example, args is a list of the form

-background backgroundColor -foreground foregroundColor

The -background, -font, -foreground, -selectbackground, and -selectforeground cell configuration options override the options of the same names set at row, column, or widget level if the specified value is not an empty string.  See the COLORS AND FONTS section for further details on these options.

DESCRIPTION

The tablelist::tablelist command creates a new window named pathName and of the class Tablelist, and makes it into a tablelist widget.  Additional options, described above, may be specified on the command line or in the option database to configure aspects of the tablelist such as its colors, font, and columns.  The tablelist::tablelist command returns its pathName argument.  At the time this command is invoked, there must not exist a window named pathName, but pathName's parent must exist.

A tablelist is a multi-column listbox and tree widget, implemented as a mega-widget, consisting of a body and a header.  The body displays a list of items, one per line.  Each item is a list of elements, which are aligned in columns.  In other words, an item is the contents of a row, and an element is the text contained in a cell.  The header consists of label widgets displaying the column titles.  The labels can be used, among others, for interactive column resizing and column-based sorting of the items, as described below.

Each cell and each header label of a tablelist widget can also contain an image, which is placed to the left or right of the text, depending on the column's alignment.  Instead of an image, a tablelist cell can also contain an embedded window, placed to the left or right of the text, just like an embedded image.

When a tablelist is used as a tree widget, one of its columns will display the tree hierarchy with the aid of indentations and expand/collapse controls.  The Tablelist package provides a great variety of tree styles controlling the look & feel of that column, and chooses the correct default style depending on the windowing system, operating system version, and tile theme.  In a tablelist used as a multi-column tree widget, every row is at the same time a tree node, having exactly one parent node and any number of child nodes.  The tree's origin is the invisible root node, which has no parent itself and whose children are the top-level nodes.

The elements of a tablelist widget can, per default, be only edited programmatically.  However, interactive editing can be enabled for individual cells and for entire columns.  Per default, the interactive cell editing uses a temporary embedded entry widget, thus making sure that all the validation facilities available for entry widgets can be used during the editing process.  A great variety of widgets from the packages BWidget, Iwidgets, combobox (by Bryan Oakley), ctext, and Mentry (or Mentry_tile), as well as Tk core text, spinbox, checkbutton, and menubutton widgets, along with tile entry, spinbox, combobox, checkbutton, and menubutton widgets are also supported as temporary embedded widgets used for cell editing.  In addition, a rich set of keyboard bindings is provided for a comfortable navigation between the editable cells.

When first created, a new tablelist widget has no items.  Items may be added, deleted, or updated using widget commands described below.  In addition, one or more items or elements may be selected.  If a tablelist widget is exporting its selection (see the -exportselection option), then it will observe the standard X11 protocols for handling the selection.  Tablelist widget selections are available as types STRING and UTF8_STRING; the value of the selection will be a text built by taking all of the rows having at least one viewable selected element, joining these elements together with tabs, and the resulting strings in turn with newlines.  If a tablelist widget that is exporting its selection is the selection owner and some other window claims ownership of the selection away from it, then the virtual event <<TablelistSelectionLost>> is generated.

It is not necessary for all the elements to be displayed in the tablelist widget at once; commands described below may be used to change the view in the window.  Tablelist widgets allow scrolling in both directions using the standard -xscrollcommand and -yscrollcommand options.  They also support scanning, as described below.

Each tablelist widget, as well as each tablelist column, row, and cell may have any number of attributes, which can be used in commands that create or manipulate tablelist widgets for particular purposes.

COLORS AND FONTS

The -background, -font, -foreground, -selectbackground, and -selectforeground options can also be specified at column, row, and cell level, by using the columnconfigure (or configcolumnlist, or configcolumns), rowconfigure (or configrowlist, or configrows), and cellconfigure (or configcelllist, or configcells) subcommands of the Tcl command associated with the tablelist widget.  Likewise, the -stripebackground and -stripeforeground options can also be specified at column level.  For this reason, a particular cell can have up to four values for one and the same color or font option.  If these values conflict, then the option specified at the highest priority level is used.  The decreasing priority order is cell, row, column, widget.

If one of these options hasn't been specified at cell, row, or column level, or if its value for that level is an empty string (this is explicitly allowed), then that option will not be used at that particular level.

USING A TABLELIST AS MULTI-COLUMN TREE WIDGET

When using a tablelist widget to display a tree hierarchy, there are a few rules to follow:

1. Do not use the the widget's -listvariable option, or at least do not add new items to the variable specified as its value, because adding new items to that list will result in inserting those list elements into the widget without respecting the tree's internal structure.
2. Make sure that the column specified by the -treecolumn option is left-aligned so the tree structure can be displayed as expected.
3. Use the insertchild(ren) or insertchildlist subcommand rather than insert or insertlist for inserting new items into the widget.

INTERACTIVE CELL EDITING

Whether or not the contents of a cell of a tablelist widget can be edited interactively, depends on the -editable option on both cell and column level.  If the cell-level option was set explicitly then its value determines the editability of the cell's contents.  Otherwise the value of the column-level option is used to decide whether the cell can be edited interactively.  From this rule it follows that you can enable interactive cell editing for a whole column by setting its -editable option to true.  To exclude some of the column's cells from interactive editing, set their -editable option to false.

The interactive cell editing is started by the editcell subcommand, which is invoked implicitly by clicking with the left mouse button into an editable cell (see the -editselectedonly option for details) or using keyboard navigation to move from one editable cell into another.  If the selection type is cell and the location cursor is in an editable cell, then the interactive editing of the active element can also be started by pressing Return or KP_Enter.  Per default, the editcell subcommand creates a temporary entry widget and embeds it into the cell whose index was passed to it as argument.  You can, however, use the -editwindow column or cell configuration option to specify another widget instead of an entry, like a Tk core text, spinbox, checkbutton, or menubutton widget, or a tile entry, spinbox, combobox, checkbutton, or menubutton, or one of the 19 currently supported widgets from the packages BWidget, Iwidgets, combobox (by Bryan Oakley), ctext, and Mentry (or Mentry_tile).  Before specifying a widget from one of these library packages as the value of the -editwindow column or cell configuration option, you must register it for interactive cell editing by invoking the corresponding tablelist::add* command.  The above-mentioned Tk core and tile widgets are automatically registered for cell editing.

In the simplest case, the text automatically inserted into the temporary embedded widget is the same as the text displayed in the cell, which in turn can be the cell's contents or the string obtained from the latter by using the -formatcommand option of the cell's column.  However, if the value of the -editstartcommand configuration option is a nonempty string, then the text displayed in the cell is passed to that command as its last argument (following the tablelist's path name as well as the cell's row and column indices), the resulting script is evaluated in the global scope, and the return value becomes the edit window's contents.  From within this script you can invoke the cancelediting subcommand, which destroys the temporary embedded widget and cancels the editing of its contents.  The main goal of this script is, however, to enable you to define validations for the editing process.  This can be done either with the aid of the options for entry validation, supported by Tk versions 8.3 and higher (see the entry reference page), or by using the widget callback package Wcb, available for Tk versions 8.0 and higher.  The Iwidgets package (available for Tk versions 8.0 or higher) provides its own validation facilities, which can equally be used if the edit window is a widget belonging to that extension.  In either case, you will need the path name of the temporary embedded widget or that of its entry or entry-like component; use the editwinpath and entrypath subcommands to get these path names.  Another purpose of the command indicated by the -editstartcommand option is to enable you to prepare the edit window in various other ways.  For example, if the latter is a combobox widget then you can set its -editable option to false or (for a tile combobox) set its state to readonly, and you will have to populate its listbox component.  In the same script, you can change some of the embedded widget's visual attributes (like its background, selection background, or selection foreground color).  (Notice that this can also be done with the aid of the Tk option database.)

The editing of the text inserted into the edit window uses the widget-specific bindings for mouse and keyboard events, with a few extensions and changes, as described in the DEFAULT BINDINGS FOR INTERACTIVE CELL EDITING section.  For example, in entry or entry-like components of the edit window, Control-i inserts a tab, Control-j inserts a newline, and if the edit window is a text or ctext widget then Return and KP_Enter insert a newline character, too.  Tab and Shift-Tab are used for navigation between the editable cells, just like Alt-Left, Alt-Right, Alt-Up, Alt-Down, Alt-Prior, Alt-Next, Alt-Home, and Alt-End (as well as Control-Home and Control-End, except in the case of a text or ctext widget).  On Mac OS Classic and Mac OS X Aqua the Command key is used instead of Alt.  The editing can be aborted with the Escape key (or by invoking the cancelediting subcommand) and terminated normally with Return or KP_Enter (together with Control for a text or ctext widget).  The bindings for the keys used for normal termination of the editing just invoke the finishediting subcommand; the latter can also be called explicitly to terminate the editing programmatically.  Normal termination is also triggered by clicking with the left mouse button anywhere in the tablelist's body, outside the cell just being edited, or moving into another editable cell by using keyboard navigation.  If the editing was started with the left mouse button, the edit window is a Tk or tile checkbutton, and the value of the -instanttoggle option is true, then the normal termination of the editing will take place automatically, without any user interaction.

When normal termination of the editing process occurs, the Tcl command associated with the tablelist widget compares the edit window's final contents to its original one.  If they are equal then the embedded widget is destroyed and the cell's original value is restored.  If the two strings are different and the value of the -editendcommand configuration option is a nonempty string, then the edit window's final text is passed to that command as its last argument (following the tablelist's path name as well as the cell's row and column indices), the resulting script is evaluated in the global scope, and the return value becomes the cell's new internal contents after destroying the temporary embedded widget.  The main goal of this script is to enable you to do a final validation of the edit window's contents.  From within this script you can invoke the rejectinput subcommand, which prevents the script's return value from becoming the cell's new contents; this subcommand also prevents the destruction of the temporary embedded widget.  Another purpose of the command indicated by the -editendcommand option is to convert the edit window's text to the cell's new internal contents, which is necessary if, due to the -formatcommand column configuration option, the cell's internal value is different from its external representation.  See the description of the -forceeditendcommand option for more about the invocation of the command mentioned above.

DRAG & DROP SUPPORT

As described in detail in the "DEFAULT AND INDIVIDUAL BINDINGS FOR THE TABLELIST BODY" section, the default binding scripts perform an automatic drag-friendly handling of the selection and of pointer movements with mouse button 1 down if the tablelist widget's -customdragsource opton was set to true or its body component was registered as a drag source for mouse button 1 via the  tkdnd::drag_source register  or the BWidget DragSite::register command.

On the other hand, you can also register a tablelist widget as a TkDND drop target, by invoking the  tkdnd::drop_target register  command.  Similarly, you can register the body component of a tablelist widget as a drop target for the drag & drop framework included in BWidget, with the aid of the DropSite::register command.  (Notice that for BWidget it is necessary to register the tablelist's body rather than the widget itself.)

The action to be triggered by a drop event over a tablelist widget might depend on the row under the mouse and the vertical position of the mouse pointer within the containing row (if any).  For example, your application might insert one or more new siblings before or after the row in question, or a child item having that row as parent, depending on the vertical position of the mouse pointer within the containing row.  Alternatively, the drop event might update the item under the mouse, using the data associated with the drag source, or perform some other, application-specific action.  In any case, you can significantly increase the user-friendliness of your application by displaying a target indicator (a horizontal gap or vertical bar) before or inside the row that will be affected by the drop.  The examples below demonstrate how to achieve this for a tablelist widget $tbl with the aid of the showtargetmark, hidetargetmark, targetmarkpos, and targetmarkpath subcommands.

TkDND EXAMPLE:  For an introduction to TkDND see the online TkDND Tutorial by Georgios Petasis.  For the following example it is strongly recommended to use TkDND version 2.7 or later, which can be downloaded from the location

http://sourceforge.net/projects/tkdnd/files/

package require tkdnd 2.7

tkdnd::drop_target register $tbl DND_Text

bind $tbl <<DropEnter>>    { onTblDropEnterOrPos %W %X %Y %a %b }
bind $tbl <<DropPosition>> { onTblDropEnterOrPos %W %X %Y %a %b }
bind $tbl <<DropLeave>>    { %W hidetargetmark }
bind $tbl <<Drop>>         { onTblDrop %W %A %D }

proc onTblDropEnterOrPos {tbl rootX rootY actions buttons} {
    global place row
    set y [expr {$rootY - [winfo rooty $tbl]}]
    foreach {place row} [$tbl targetmarkpos $y] {}

    if {some_optional_excluding_condition} {
        $tbl hidetargetmark
        return refuse_drop
    } else {
        $tbl showtargetmark $place $row
        return copy ;# for example
    }
}

proc onTblDrop {tbl action data} {
    handleTblDrop $tbl $data
    return $action
}

BWidget EXAMPLE:  For an introduction to drag & drop with BWidget see the online tutorial BWidget example: Drag and Drop Demo by Kevin Walzer.  The DropSite::* commands used in the following example are described in the "DropSite" reference page included in the BWidget distribution.  When using this drag & drop framework, you will have to register not only the tablelist's body, but also its target indicator as drop sites, to make sure that the target mark won't get hidden by moving the mouse cursor over it with mouse button 1 down:

package require BWidget

foreach w [list [$tbl bodypath] [$tbl targetmarkpath]] {
    DropSite::register $w -dropovercmd tblDropOverCmd -dropcmd tblDropCmd \
        -droptypes ...
}

proc tblDropOverCmd {dropTarget dragSrc event rootX rootY op dataType data} {
    # $event may be "enter", "motion", or "leave"
    set tbl [tablelist::getTablelistPath $dropTarget]
    if {[string equal $event "leave"]} {
        set newWidget [winfo containing -displayof $dropTarget $rootX $rootY]
        if {![string equal $newWidget [$tbl targetmarkpath]] &&
            ![string equal $newWidget [$tbl bodypath]]} {
            $tbl hidetargetmark
            return 2 ;# refuse the drop and re-invoke the callback on motion events
        }
    }
    
    global place row
    set y [expr {$rootY - [winfo rooty $tbl]}]
    foreach {place row} [$tbl targetmarkpos $y] {}

    if {some_optional_excluding_condition} {
        $tbl hidetargetmark
        DropSite::setcursor dot
        return 2 ;# refuse the drop and re-invoke the callback on motion events
    } else {
        $tbl showtargetmark $place $row
        DropSite::setcursor based_arrow_down
        return 3 ;# accept the drop and re-invoke the callback on motion events
    }
}

proc tblDropCmd {dropTarget dragSrc rootX rootY op dataType data} {
    set tbl [tablelist::getTablelistPath $dropTarget]
    handleTblDrop $tbl $data
    return ""
}

In both examples above, you can restrict the type of the supported drop target indicator by appending the argument -horizontal or -vertical to the targetmarkpos invocation.  When passing the -vertical option, you will have to extend the if statement following that subcommand invocation as follows:

    if {$row < 0 || some_optional_excluding_condition} {

Both examples above invoke the handleTblDrop procedure, implemented as follows:

proc handleTblDrop {tbl data} {
    $tbl hidetargetmark
    global place row

    if {[string equal $place "before"]}
        if {$row < [$tbl size]} {
            # Insert a sibling of the item indicated by $row, built from $data
            # (For a flat tablelist replace the following 2 lines with
            # "set key [$tbl insert $row ...]")
            set key [$tbl insertchild [$tbl parentkey $row] \
                     [$tbl childindex $row] ...]
        } else {
            # Append a top-level item built from $data
            # (For a flat tablelist replace the following line with
            # "set key [$tbl insert end ...]")
            set key [$tbl insertchild root end ...]
        }
    } else { ;# "inside"
        # Insert a child of the item indicated by $row, built from $data
        # Alternatively, update the item indicated by $row, using $data
        set key [$tbl insertchild $row 0 ...]
        $tbl expand $row -partly
    }

    . . .
}

VIRTUAL EVENTS

Tablelist defines a few virtual events for the purposes of notification, and makes use of the  event generate  command to send them to the tablelist widget in question (whose path name will be the value of the %W substitution in event scripts).  When using Tk 8.5 or above, most of these virtual events are generated with their -data option set to an appropriate event-specific value.  Binding scripts can access this user data as the value of the %d substitution.  Unfortunately, the -data option to  event generate  was not supported by Tk versions earlier than 8.5.  For these Tk versions, the  event generate  command will be invoked without the -data option.

The virtual events generated by Tablelist are as follows:

Virtual Event	Description	User Data (for Tk 8.5 or Later)
<<TablelistCellRestored>>	Generated by the cancelediting subcommand.	A list consisting of the numerical row and column indices of the cell whose content is being edited.
<<TablelistCellUpdated>>	Generated by the finishediting subcommand.	A list consisting of the numerical row and column indices of the cell whose content is being edited.
<<TablelistColHiddenStateChanged>>	Generated whenever the hidden state of one or more columns is toggled by using the columnconfigure, configcolumnlist, configcolumns, or togglecolumnhide subcommand.	A list consisting of the numerical indices of the columns whose hidden state changed.
<<TablelistColumnMoved>>	Generated by the default bindings whenever a column is moved interactively to a new position.	A list of length 4, whose first two elements are identical to the two arguments passed to the movecolumn subcommand invoked for moving the source column to the target position, and the last two elements are the corresponding column names, retrieved with the aid of the  columncget ... -name  subcommand.
<<TablelistColumnResized>>	Generated by the default bindings whenever a column is resized interactively.	The numerical index of the column that was resized.
<<TablelistColumnSorted>>	Generated by the tablelist::sortByColumn command.	A list consisting of the numerical column index and the sort order (decreasing or increasing).
<<TablelistColumnsSorted>>	Generated by the tablelist::addToSortColumns command.	A list consisting of the two lists passed to the sortbycolumnlist subcommand, invoked to perform the multi-column sorting.
<<TablelistRowHiddenStateChanged>>	Generated whenever the hidden state of one or more rows is toggled by using the rowconfigure, configrowlist, configrows, or togglerowhide subcommand.	A list consisting of the numerical indices of the rows whose hidden state changed.
<<TablelistRowMoved>>	Generated by the default bindings whenever a row is moved interactively to a new position.	A list of length 3, whose elements are derived from the arguments passed to the second form of the move subcommand invoked for moving the source row to the target position.  The first list element will be the full key corresponding to the first argument, the second one will be root or the full key corresponding to the second argument, and the third list element will be identical to the third argument passed to the move subcommand.
<<TablelistSelect>>, <<ListboxSelect>>	Generated by the default bindings whenever the selection changes in the tablelist widget due to user interaction.	Not used.
<<TablelistSelectionLost>>	Sent to the tablelist widget whenever it loses ownership of the PRIMARY selection.	Not used.

ROW INDICES

Many of the widget commands for tablelist widgets take one or more row indices as arguments.  A row index specifies a particular item of the tablelist, in any of the following ways:

number	Specifies the item as a numerical index, where 0 corresponds to the first item in the tablelist.
knumber	Specifies the item by its full key, composed of the letter k and the sequence number associated with the item.  You can use the getkeys widget command to get this sequence number, or the getfullkeys widget command to retrieve the full key.  In addition, the insert, insertlist, insertchildren, and insertchildlist subcommands return the list of full keys associated with the items just inserted.
active	Indicates the item containing the element that has the location cursor.  Depending on the selection type, this item as a whole or just its element having the location cursor will be displayed according to the value of the -activestyle configuration option when the tablelist has the keyboard focus.  This item is specified with the activate widget command or as the row containing the element specified with the activatecell widget command.
anchor	Indicates the anchor point for the row selection, which is set with the  selection anchor  widget command, or the row containing the anchor point for the cell selection, which is set with the  cellselection anchor  widget command.
end	Indicates the end of the tablelist.  For most commands this refers to the last item in the tablelist, but for a few commands such as index, insert, insertlist, and  showtargetmark before,  as well as for the target of the move command it refers to the item just after the last one.
last	Indicates the last item of the tablelist.
top	Indicates the topmost item visible in the tablelist window.
bottom	Indicates the bottommost item visible in the tablelist window.
@x,y	Indicates the item that covers the point in the tablelist window specified by x and y (in pixel coordinates).  If no item covers that point, then the closest item to that point is used.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).
name	Specifies the row by the value of its -name configuration option.  name must be different from all the above row indices, and should be unique (if several rows have the same name then this value is considered to indicate the first matching row).

In the widget command descriptions below, arguments named index, first, last, sourceIndex, and targetIndex always contain row indices in one of the above forms.

NODE INDICES

Many of the widget commands for tablelist widgets take a node index as argument.  A node index is either the word root, denoting the invisible root node of a tablelist used as a tree widget, or a regular row index in one of the above forms, specifying a particular item of the tablelist.

In the widget command descriptions below, arguments named nodeIndex, parentNodeIndex, and targetParentNodeIndex always contain node indices.

COLUMN INDICES

Many of the widget commands for tablelist widgets take one or more column indices as arguments.  A column index specifies a particular column of the tablelist, in any of the following ways:

number	Specifies the column as a numerical index, where 0 corresponds to the first column in the tablelist.
active	Indicates the column containing the element that has the location cursor.  If the selection type is cell then this element will be displayed according to the value of the -activestyle configuration option when the tablelist has the keyboard focus.  This element is specified with the activatecell widget command.
anchor	Indicates the column containing the anchor point for the cell selection, which is set with the  cellselection anchor  widget command.
end	Indicates the last column of the tablelist, except for the commands insertcolumns and insertcolumnlist, as well as for the target of the movecolumn command, in which cases it refers to the column just after the last one.
last	Indicates the last column of the tablelist.
left	Indicates the leftmost column visible in the tablelist window.
right	Indicates the rightmost column visible in the tablelist window.
@x,y	Indicates the column that covers the point in the tablelist window specified by x and y (in pixel coordinates).  If no column covers that point, then the closest column to that point is used.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).
name	Specifies the column by the value of its -name configuration option.  name must be different from all the above column indices, and should be unique (if several columns have the same name then this value is considered to indicate the first matching column).

In the widget command descriptions below, arguments named columnIndex, firstColumn, lastColumn, sourceColumn, and targetColumn always contain column indices in one of the above forms.

CELL INDICES

Many of the widget commands for tablelist widgets take one or more cell indices as arguments.  A cell index specifies a particular cell of the tablelist, in any of the following ways:

row,col	Indicates the cell having the row index row and column index col.  row may be a number, a full key (of the form knumber), active, anchor, end (where the latter indicates the last row in the tablelist), last, top, bottom, or a row name.  col may be a number, active, anchor, end, last, left, right, or a column name.
active	Indicates the element that has the location cursor.  If the selection type is cell then this element will be displayed according to the value of the -activestyle configuration option when the tablelist has the keyboard focus.  This element is specified with the activatecell widget command.
anchor	Indicates the anchor point for the cell selection, which is set with the  cellselection anchor  widget command.
end	Indicates the last cell in the last row of the tablelist.
last	Indicates the last cell in the last row of the tablelist (same as end).
@x,y	Indicates the cell that covers the point in the tablelist window specified by x and y (in pixel coordinates).  If no cell covers that point, then the closest cell to that point is used.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).

In the widget command descriptions below, arguments named cellIndex, firstCell, and lastCell always contain cell indices in one of the above forms.

WIDGET COMMAND

The tablelist::tablelist command creates a new Tcl command whose name is pathName.  This command may be used to invoke various operations on the widget.  It has the following general form:

pathName option ?arg arg ...?

option and the args determine the exact behavior of the command.  The following commands are possible for tablelist widgets:

pathName activate index

Sets the active item to the one indicated by index if the tablelist's state is not disabled.  If index is outside the range of items in the tablelist or it refers to a non-viewable item then the closest viewable item is activated.  The active item is drawn as specified by the -activestyle configuration option when the widget has the input focus and the selection type is row.  Its index may be retrieved with the index active.  Returns an empty string.

pathName activatecell cellIndex

Sets the active element to the one indicated by cellIndex if the tablelist's state is not disabled.  If cellIndex is outside the range of elements in the tablelist or it refers to a non-viewable element, then the closest viewable element is activated.  The active element is drawn as specified by the -activestyle configuration option when the widget has the input focus and the selection type is cell.  Its index may be retrieved with the cell index active.  Returns an empty string.

pathName applysorting itemList

Sorts the elements of itemList according to the parameters of the most recent sort, sortbycolumn, or sortbycolumnlist invocation and returns a new list in sorted order.  The itemList argument is supposed to be a well-formed list of lists, where the length of each sublist is expected to be no less than the number of columns of the tablelist widget.  If the items haven't been sorted at all, or the sort information was reset by invoking resetsortinfo, then no sorting takes place and the return value will be identical to itemList.

REMARK:  This subcommand is usually invoked from within the command specified as the value of the -expandcommand configuration option, for a tablelist row whose children have not yet been inserted into the widget.  The most efficient way to insert the child items in the correct order is to add them to a list, sort this list via applysorting, and then insert the items of the sorted list by invoking the insertchildlist subcommand.

pathName attrib ?name? ?value name value ...?

Queries or modifies the attributes of the widget.  If no name is specified, the command returns a list of pairs, each of which contains the name and the value of an attribute for pathName.  If name is specified with no value, then the command returns the value of the one named attribute, or an empty string if no corresponding value exists (you can use the hasattrib subcommand to distinguish this case from the one that the value of an existing attribute is an empty string).  If one or more name-value pairs are specified, then the command sets the given widget attribute(s) to the given value(s); in this case the return value is an empty string.  name may be an arbitrary string.

pathName bbox index

Returns a list of four numbers describing the bounding box of the row given by index.  The first two elements of the list give the x and y coordinates of the upper-left corner of the screen area covered by the row (specified in pixels relative to the widget) and the last two elements give the width and height of the area, in pixels.  If no part of the row given by index is visible on the screen, or if index refers to a non-existent row, then the result is an empty string; if the row is partially visible, the result gives the full area of the row, including any parts that are not visible.

pathName bodypath

Returns the path name of the body component of the widget.

pathName bodytag

Returns the name of a binding tag whose name depends on the path name of the tablelist widget and which is associated with the tablelist's body, the separators, as well as the message and label widgets used for displaying multi-line elements and images, respectively.  This binding tag precedes the tag TablelistBody in the list of binding tags of the tablelist descendants mentioned above, and is designed to be used when defining individual binding scripts for tablelist widgets.  The main advantage of using this tag instead of the path name of the tablelist's body is that it enables you to write event handling scripts that are valid not only for the tablelist's body but also for the separators, multi-line cells, and embedded images.

pathName canceledediting

Returns the value 1 if the most recent interactive cell editing was canceled (by the user or programmatically, with the aid of the cancelediting subcommand) and 0 otherwise.

REMARK:  By invoking this subcommand from within the Tcl command specified by the -editendcommand configuration option, you can find out whether the editing session was canceled or terminated normally.  Remember that this Tcl command will be invoked in both cases mentioned above if the -forceeditendcommand option was set to true.

pathName cancelediting

This subcommand cancels the interactive editing of the contents of the cell whose index was passed to the editcell subcommand, destroys the temporary widget embedded into the cell, and restores the original cell contents.  This command enables you to cancel the interactive cell editing from within the Tcl command specified by the -editstartcommand configuration option if that pre-edit callback encounters an error when preparing the text to be inserted into the edit window.  The command is also invoked implicitly by pressing the Escape key when a cell is being edited.  The return value is an empty string.  Immediately before returning this value, the command generates the virtual event <<TablelistCellRestored>>.  For Tk versions 8.5 or higher, this virtual event is generated with its -data option set to a list consisting of the numerical row and column indices of the cell whose content is being edited.  If no cell was being edited when the command was invoked then an empty string is returned without generating a virtual event.

pathName cellattrib cellIndex ?name? ?value name value ...?

Queries or modifies the attributes of the cell given by cellIndex.  If no name is specified, the command returns a list of pairs, each of which contains the name and the value of an attribute for the cell.  If name is specified with no value, then the command returns the value of the one named cell attribute, or an empty string if no corresponding value exists (you can use the hascellattrib subcommand to distinguish this case from the one that the value of an existing cell attribute is an empty string).  If one or more name-value pairs are specified, then the command sets the given cell attribute(s) to the given value(s); in this case the return value is an empty string.  name may be an arbitrary string.

pathName cellbbox cellIndex

Returns a list of four numbers describing the bounding box of the cell given by cellIndex.  The first two elements of the list give the x and y coordinates of the upper-left corner of the screen area covered by the cell (specified in pixels relative to the widget) and the last two elements give the width and height of the area, in pixels.  If no part of the cell given by cellIndex is visible on the screen, or if cellIndex refers to a non-existent cell, then the result is an empty string; if the cell is partially visible, the result gives the full area of the cell, including any parts that are not visible.

pathName cellcget cellIndex option

Returns the current value of the cell configuration option given by option for the cell specified by cellIndex.  option may have any of the values accepted by the cellconfigure command.

pathName cellconfigure cellIndex ?option? value option value ...?

Queries or modifies the configuration options of the cell given by cellIndex.  If no option is specified, the command returns a list describing all of the available options for the cell (see Tk_ConfigureInfo for information on the format of this list).  If option is specified with no value, then the command returns a list describing the one named option (this list will be identical to the corresponding sublist of the value returned if no option is specified).  If one or more option-value pairs are specified, then the command modifies the given cell option(s) to have the given value(s); in this case the return value is an empty string.  option may have any of the values described in the CELL CONFIGURATION OPTIONS section.

pathName cellindex cellIndex

Returns the canonical cell index value that corresponds to cellIndex, in the form row,col, where row and col are integers.

pathName cellselection option args

This command is used to adjust the cell selection within a tablelist widget.  It has several forms, depending on option:

pathName cellselection anchor cellIndex

Sets the cell selection anchor to the element given by cellIndex.  If cellIndex refers to a nonexistent or non-viewable element, then the closest viewable element is used.  The cell selection anchor is the end of the cell selection that is fixed while dragging out a cell selection with the mouse if the selection type is cell.  The cell index anchor may be used to refer to the anchor element.

pathName cellselection clear firstCell lastCell
pathName cellselection clear cellIndexList

If any of the elements between firstCell and lastCell (inclusive) or corresponding to the cell indices specified by the list cellIndexList are selected, they are deselected.  The selection state is not changed for elements outside the range given in the first form of the command or different from those specified by the cell index list given in its second form.

pathName cellselection includes cellIndex

Returns 1 if the element indicated by cellIndex is currently selected, 0 if it isn't.

pathName cellselection set firstCell lastCell
pathName cellselection set cellIndexList

Selects all of the selectable elements in the range between firstCell and lastCell, inclusive, or corresponding to the indices specified by the list cellIndexList, without affecting the selection state of any other elements.  An element is viewed as selectable if and only if the value of the -selectable option of the row containing it is true.

If the tablelist's state is disabled and option is different from includes then the command just returns an empty string, without performing any of the above actions.

pathName cget option

Returns the current value of the configuration option given by option, which may have any of the values accepted by the tablelist::tablelist command.

pathName childcount nodeIndex

Returns the number of children of the tree node indicated by nodeIndex.  If this argument is specified as root then the return value will be the number of top-level items of the tablelist widget.

pathName childindex index

Returns the numerical index of the row given by index in the list of children of its parent node.

pathName childkeys nodeIndex

Returns the list of full keys of the children of the tree node indicated by nodeIndex.  If this argument is specified as root then the return value will be the list of full keys of the top-level items contained in the tablelist widget.

pathName collapse index ?-fully|-partly?

This subcommand collapses the specified row of a tablelist used as a tree widget, i.e., elides all its descendants.  The optional argument -fully (which is the default) indicates that the command will be performed recursively, i.e., all of the descendants of the node specified by index will be collapsed, so that a subsequent invocation of the non-recursive version of the expand(all) subcommand will only display the row's children but no further descendants of it.  The -partly option (which is used by the default bindings) restricts the operation to just one hierarchy level, implying that by a subsequent invocation of the non-recursive version of the expand(all) subcommand exactly the same descendants will be displayed again that were visible prior to collapsing the row.

Before hiding the descendants of a row, the command specified as the value of the -collapsecommand option (if any) is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this subcommand is not supported for Tk versions earlier than 8.3.

pathName collapseall ?-fully|-partly?

This subcommand collapses all top-level rows of a tablelist used as a tree widget, i.e., elides all their descendants.  The optional argument -fully (which is the default) indicates that the command will be performed recursively, i.e., all of the descendants of the top-level nodes will be collapsed, so that a subsequent invocation of the non-recursive version of the expandall subcommand will only display ther children but no further descendants of them.  The -partly option restricts the operation to just one hierarchy level, implying that by a subsequent invocation of the non-recursive version of the expandall subcommand exactly the same items will be displayed again that were visible prior to collapsing the top-level ones.

Before hiding the descendants of a row, the command specified as the value of the -collapsecommand option (if any) is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this subcommand is not supported for Tk versions earlier than 8.3.

pathName columnattrib columnIndex ?name? ?value name value ...?

Queries or modifies the attributes of the column given by columnIndex.  If no name is specified, the command returns a list of pairs, each of which contains the name and the value of an attribute for the column.  If name is specified with no value, then the command returns the value of the one named column attribute, or an empty string if no corresponding value exists (you can use the hascolumnattrib subcommand to distinguish this case from the one that the value of an existing column attribute is an empty string).  If one or more name-value pairs are specified, then the command sets the given column attribute(s) to the given value(s); in this case the return value is an empty string.  name may be an arbitrary string.

pathName columncget columnIndex option

Returns the current value of the column configuration option given by option for the column specified by columnIndex.  option may have any of the values accepted by the columnconfigure command.

pathName columnconfigure columnIndex ?option? ?value option value ...?

Queries or modifies the configuration options of the column given by columnIndex.  If no option is specified, the command returns a list describing all of the available options for the column (see Tk_ConfigureInfo for information on the format of this list).  If option is specified with no value, then the command returns a list describing the one named option (this list will be identical to the corresponding sublist of the value returned if no option is specified).  If one or more option-value pairs are specified, then the command modifies the given column option(s) to have the given value(s); in this case the return value is an empty string.  option may have any of the values described in the COLUMN CONFIGURATION OPTIONS section.

pathName columncount

Returns the number of columns in the tablelist widget.

pathName columnindex columnIndex

Returns the integer column index value that corresponds to columnIndex.

pathName columnwidth columnIndex ?-requested|-stretched|-total?

Returns the current width in pixels of the column specified by columnIndex.  If the optional argument is -requested (which is the default) then the return value is the number of pixels corresponding to the column width specified by the -columns widget or -width column configuration option (possibly overridden by interactive column resizing); if the column width was specified as 0 (and was not changed by interactive column resizing) then the return value is the actual number of pixels corresponding to the widest viewable element of the column, including its header.  With the -stretched option, the command returns the column width obtained by increasing the value described above by the number of additional pixels that might have been added to the requested column width by a stretch operation (see the -stretch widget and -stretchable column configuration options).  Finally, if the optional argument is -total then the return value is the stretched column width increased by the number of pixels corresponding to the left and right margins within the column; this value equals the width of the header label if the tablelist widget is mapped.

pathName configcelllist cellConfigSpecList

For each cell index, option, and value specified by the list cellConfigSpecList, the command modifies the given option of the given cell to have the given value.  The argument cellConfigSpecList must be a list of the form

cellIndex option value cellIndex option value ...

where each option may have any of the values described in the CELL CONFIGURATION OPTIONS section.  The return value is an empty string.

This command has the same effect as

eval [list pathName configcells] cellConfigSpecList

but it is more efficient and easier to use.

pathName configcells ?cellIndex option value cellIndex option value ...?

For each cellIndex, option, and value, the command modifies the given option of the given cell to have the given value.  Each option may have any of the values described in the CELL CONFIGURATION OPTIONS section.  The return value is an empty string.

pathName configcolumnlist columnConfigSpecList

For each column index, option, and value specified by the list columnConfigSpecList, the command modifies the given option of the given column to have the given value.  The argument columnConfigSpecList must be a list of the form

columnIndex option value columnIndex option value ...

where each option may have any of the values described in the COLUMN CONFIGURATION OPTIONS section.  The return value is an empty string.

This command has the same effect as

eval [list pathName configcolumns] columnConfigSpecList

but it is more efficient and easier to use.

pathName configcolumns ?columnIndex option value columnIndex option value ...?

For each columnIndex, option, and value, the command modifies the given option of the given column to have the given value.  Each option may have any of the values described in the COLUMN CONFIGURATION OPTIONS section.  The return value is an empty string.

pathName configrowlist rowConfigSpecList

For each row index, option, and value specified by the list rowConfigSpecList, the command modifies the given option of the given row to have the given value.  The argument rowConfigSpecList must be a list of the form

index option value index option value ...

where each option may have any of the values described in the ROW CONFIGURATION OPTIONS section.  The return value is an empty string.

This command has the same effect as

eval [list pathName configrows] rowConfigSpecList

but it is more efficient and easier to use.

pathName configrows ?index option value index option value ...?

For each index, option, and value, the command modifies the given option of the given row to have the given value.  Each option may have any of the values described in the ROW CONFIGURATION OPTIONS section.  The return value is an empty string.

pathName configure ?option? ?value option value ...?

Queries or modifies the configuration options of the widget.  If no option is specified, the command returns a list describing all of the available options for pathName (see Tk_ConfigureInfo for information on the format of this list).  If option is specified with no value, then the command returns a list describing the one named option (this list will be identical to the corresponding sublist of the value returned if no option is specified).  If one or more option-value pairs are specified, then the command modifies the given widget option(s) to have the given value(s); in this case the return value is an empty string.  option may have any of the values accepted by the tablelist::tablelist command.

pathName containing y

Given a y-coordinate within the tablelist window, this command returns the index of the tablelist item containing that y-coordinate.  If no corresponding item is found then the return value is -1.  The coordinate y is expected to be relative to the tablelist window itself (not its body component).

pathName containingcell x y

Given an x- and a y-coordinate within the tablelist window, this command returns the index of the tablelist cell containing the point having these coordinates.  If no corresponding cell is found then the row or column component (or both) of the return value is -1.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).

pathName containingcolumn x

Given an x-coordinate within the tablelist window, this command returns the index of the tablelist column containing that x-coordinate.  If no corresponding column is found then the return value is -1.  The coordinate x is expected to be relative to the tablelist window itself (not its body component).

pathName cornerlabelpath

Returns the path name of the label widget contained in the corner frame (see the cornerpath subcommand.  When using Tablelist_tile, the return value will be a ttk::label widget of the same style as the header labels.  The global visual options set for the header labels are automatically applied to this (ttk::)label widget, too.

pathName cornerpath

Returns the path name of a frame widget that is a sibling of the tablelist and is automatically created and destroyed together with the latter's header.  This frame is called the corner frame, because it is designed to be shown to the right of the tablelist widget's top-right corner when managing the vertical scrollbar on Mac OS X Aqua (see below).  The corner frame contains a label or ttk::label widget, depending on whether the package Tablelist or Tablelist_tile is being used (see the cornerlabelpath subcommand).  Whenever the height of the tablelist's header changes, the height of the corner frame as well as that of the above-mentioned (ttk::)label widget is updated accordingly.

REMARK:  This subcommand enables you to manage the vertical scrollbar (if any) to appear below the tablelist widget's header, thus respecting the native look & feel on Mac OS X Aqua.  As shown in the following example, it is recommended to always create a scrolled tablelist along with the scrollbar(s) as children of a (ttk::)frame widget:

# Add some entries to the Tk option database
set winSys [tk windowingsystem]
if {[string compare $winSys "x11"] != 0} {
    option add *ScrollArea.borderWidth                  1
    option add *ScrollArea.relief                       sunken
    option add *ScrollArea.Tablelist.borderWidth        0
    option add *ScrollArea.Tablelist.highlightThickness 0
}

. . .

# Create the tablelist and the scrollbars as
# children of a frame of class ScrollArea
set frm [(ttk::)frame ... -class ScrollArea]
set tbl $frm.tbl
set vsb $frm.vsb
set hsb $frm.hsb
tablelist::tablelist $tbl ... \
    -xscrollcommand [list $hsb set] -yscrollcommand [list $vsb set]
scrollbar $vsb -orient vertical   -command [list $tbl yview]
scrollbar $hsb -orient horizontal -command [list $tbl xview]

. . .

# Manage the widgets within the frame
grid $tbl -row 0 -rowspan 2 -column 0 -sticky news
if {[string compare $winSys "aqua"] == 0} {
    grid [$tbl cornerpath] -row 0 -column 1 -sticky ew
    grid $vsb              -row 1 -column 1 -sticky ns
} else {
    grid $vsb -row 0 -rowspan 2 -column 1 -sticky ns
}
grid $hsb -row 2 -column 0 -sticky ew
grid rowconfigure    $frm 1 -weight 1
grid columnconfigure $frm 0 -weight 1

# Manage the frame
pack $frm -expand yes -fill both

pathName curcellselection

Returns a list containing the canonical indices (of the form row,col, where row and col are numbers) of all of the elements in the tablelist that are currently selected.  If there are no such elements in the tablelist then an empty string is returned.

pathName curselection

Returns a list containing the numerical indices of all of the items in the tablelist that contain at least one selected element.  If there are no such items in the tablelist then an empty string is returned.

pathName delete first last
pathName delete indexList

Deletes one or more items of the tablelist if its state is not disabled.  In the first form of the command, first and last are indices specifying the first and last items in the range to delete.  The command's second form accepts a list indexList of indices specifying the items to be deleted.  In both cases, all descendants of the specified items will be deleted, too.  The return value is an empty string.

pathName deletecolumns firstColumn lastColumn
pathName deletecolumns columnIndexList

Deletes one or more columns of the tablelist if its state is not disabled.  In the first form of the command, firstColumn and lastColumn are indices specifying the first and last columns in the range to delete.  The command's second form accepts a list columnIndexList of indices specifying the columns to be deleted.  Returns an empty string.

pathName depth nodeIndex

Returns the number of steps from the tree node indicated by nodeIndex to the root node.  The latter's depth is 0, that of the top-level rows is 1, and so on.

pathName descendantcount nodeIndex

Returns the number of descendants of the tree node indicated by nodeIndex.  If this argument is specified as root then the return value will be the number of items of the tablelist widget.

pathName editcell cellIndex

Starts the interactive editing of the cell's contents if the tablelist's state is not disabled and the cell is viewable and editable.  Returns an empty string.  See the INTERACTIVE CELL EDITING section for details on editablity and on the editing process.

pathName editinfo

Returns a three-element list containing information about the tablelist cell whose content is being edited.  The first element of the list is the full key (of the form knumber) associated with the item containing the tablelist element that is being edited.  The second and third elements are the cell's row and column indices.  If currently no cell is being edited then the return value is the list  {"" -1 -1}.

pathName editwinpath

Returns the path name of the temporary embedded widget used for interactive cell editing, created by the editcell subcommand.  If no cell is currently being edited then the return value is an empty string.  This subcommand enables you to access the edit window from within the commands specified by the -editstartcommand and -editendcommand configuration options.

pathName editwintag

Returns the name of a binding tag whose name depends on the path name of the tablelist widget and which is associated with some of the components of the temporary embedded widget used for interactive cell editing.  These components depend on the edit window: the widget itself in case of a Tk or tile checkbutton or menubutton; the edit window's entry children in case of a mentry widget; the only entry or entry-like component of the edit window in all other cases (see also the entrypath subcommand).  This binding tag precedes the tag TablelistEdit in the list of binding tags of the edit window components mentioned above, and is designed to be used when defining individual binding scripts for controlling the interactive cell editing.

For example, the following command will replace the standard behavior of the Return key during cell editing in the tablelist widget .tbl with that of the Tab key:

bind [.tbl editwintag] <Return> "[bind TablelistEdit <Tab>]; break"

pathName entrypath

Returns the path name of the entry or entry-like component of the temporary embedded widget used for interactive cell editing, created by the editcell subcommand.  If no cell is currently being edited or the editing is taking place with the aid of a Tk or tile checkbutton, Tk or tile menubutton, or mentry widget, then the return value is an empty string; otherwise it is the path name of a Tk or tile entry, text or ctext, Tk or tile spinbox, or BWidget Entry widget, which can be the edit window itself or one of its descendants.  This subcommand enables you to access the entry or entry-like component of the temporary embedded widget from within the commands specified by the -editstartcommand and -editendcommand configuration options.

pathName expand index ?-fully|-partly?

This subcommand expands the specified row of a tablelist used as a tree widget, i.e., makes all its children visible.  The optional argument -fully (which is the default) indicates that the command will be performed recursively, i.e., all of the descendants of the node specified by index will be displayed.  The -partly option (which is used by the default bindings) restricts the operation to just one hierarchy level, indicating that only the children of the specified node will be displayed, without changing the expanded/collapsed state of the child nodes.

Before displaying the children of a row, the command specified as the value of the -expandcommand option (if any) is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope.  This enables you to insert a tree node's children on demand, just before expanding it.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this subcommand is not supported for Tk versions earlier than 8.3.

pathName expandall ?-fully|-partly?

This subcommand expands all top-level rows of a tablelist used as a tree widget, i.e., makes all their children visible.  The optional argument -fully (which is the default) indicates that the command will be performed recursively, i.e., all of the descendants of the top-level nodes will be displayed.  The -partly option restricts the operation to just one hierarchy level, indicating that only the children of the top-level nodes will be displayed, without changing the expanded/collapsed state of the child nodes.

Before displaying the children of a row, the command specified as the value of the -expandcommand option (if any) is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope.  This enables you to insert a tree node's children on demand, just before expanding it.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this subcommand is not supported for Tk versions earlier than 8.3.

pathName expandedkeys

Returns the list of full keys of the expanded items.

pathName fillcolumn columnIndex text

Sets all the elements of the specified column to the value text.

pathName findcolumnname name

Returns the index of the first column whose name (given by the -name column configuration option) equals name.  If no column has the specified name then the return value is -1.

pathName findrowname name ?-descend? ?-parent nodeIndex?

Returns the row index of the first child of the tree node indicated by nodeIndex whose name (given by the -name row configuration option) equals name.  The -descend option extends the search to all descendants of the tree node given by the -parent option.  The default is to restrict the search to the parent node's children only.  The default parent is root.  When searching for the given name, the items will be visited in the order of their row indices.  If none of the relevant rows has the specified name then the return value is -1.

pathName finishediting

This subcommand attempts to terminate the interactive editing of the contents of the cell whose index was passed to the editcell subcommand by destroying the temporary widget embedded into the cell and updating the cell's contents.  The exact steps involved are as follows:  First, the widget's final text is compared to its original one.  If they are equal then the edit window is destroyed and the cell's original contents are restored.  If the two strings are different and the value of the -editendcommand configuration option is a nonempty string, then the widget's final text is passed to that command as its last argument (following the tablelist's path name as well as the cell's row and column indices), the resulting script is evaluated in the global scope, and the return value becomes the cell's new contents after destroying the edit window.  However, if from within this script the rejectinput subcommand was invoked then the cell's value is not changed and the embedded widget remains displayed in the cell; in this case the command returns the boolean value 0.  In all the other cases, the return value is 1.  Immediately before returning the value 1, the command generates the virtual event <<TablelistCellUpdated>>.  For Tk versions 8.5 or higher, this virtual event is generated with its -data option set to a list consisting of the numerical row and column indices of the cell whose content is being edited.  If no cell was being edited when the command was invoked then the same value 1 is returned but no virtual event is generated.

This subcommand is called implicitly by pressing Return or KP_Enter (together with Control if the edit window is a text or ctext widget) when editing a cell, or by clicking with the left mouse button anywhere in the tablelist's body, outside the cell just being edited, or moving into another editable cell by using keyboard navigation.  If the editing was started with the left mouse button, the edit window is a Tk or tile checkbutton, and the value of the -instanttoggle option is true, then this subcommand will be invoked automatically, without any user interaction.

REMARK:  There are also situations where an explicit invocation of this subcommand is needed, in order to make sure that the cell just being edited gets updated with the text entered by the user.  For example, if a tablelist widget is part of a dialog used for editing some data, then the command associated with the button designed to accept the data should call this subcommand, because otherwise, if the button is pressed during interactive cell editing then the text entered into the edit window will get lost.

pathName formatinfo

This command is designed to be invoked from within a Tcl command specified as the value of the -formatcommand column configuration option.  It returns a three-element list containing information about the tablelist cell whose content is being formatted with the aid of that command.  The first element of the list is the full key (of the form knumber) associated with the item containing the tablelist element that is being formatted.  The second and third elements are the cell's row and column indices.

REMARK:  This subcommand is needed in cases where the result of the formatting should depend on the cell's row.  To be able to use it, specify the value of the -formatcommand column configuration option as  [list formatCommand pathName],  like in the following example:

.tbl columnconfigure 1 -formatcommand [list formatValue .tbl]

proc formatValue {tbl val} {
    # Get information about the cell whose content is being formatted
    foreach {key row col} [$tbl formatinfo] {}

    # Return a string depending on $val and $row (or $key)
    . . .
}

pathName get first last
pathName get indexList

The first form of the command returns a list whose elements are all of the tablelist items (i.e., row contents) between first and last, inclusive.  The value returned by the second form depends on the number of elements in the list indexList: if the latter contains exactly one index then the return value is the tablelist item indicated by that index (or an empty string if the index refers to a non-existent item); otherwise the command returns the list of all of the tablelist items corresponding to the indices specified by indexList.

pathName getcells firstCell lastCell
pathName getcells cellIndexList

The first form of the command returns a list whose elements are all of the tablelist elements (i.e., cell contents) between firstCell and lastCell, inclusive.  The value returned by the second form depends on the number of elements in the list cellIndexList: if the latter contains exactly one cell index then the return value is the tablelist element indicated by that cell index; otherwise the command returns the list of all of the tablelist elements corresponding to the cell indices specified by cellIndexList.

pathName getcolumns firstColumn lastColumn
pathName getcolumns columnIndexList

The first form of the command returns a list whose elements are lists themselves, where each of the sublists corresponds to exactly one column between firstColumn and lastColumn, inclusive, and consists of all of the tablelist elements contained in that column.  The value returned by the second form depends on the number of elements in the list columnIndexList: if the latter contains exactly one column index then the return value is a list consisting of all of the tablelist elements contained in the column indicated by that column index; otherwise the command returns a list whose elements are lists themselves, where each of the sublists corresponds to exactly one column index in columnIndexList and consists of all of the tablelist elements contained in that column.

pathName getformatted first last
pathName getformatted indexList

The first form of the command returns a list whose elements are all of the formatted tablelist items (i.e., formatted row contents) between first and last, inclusive.  The value returned by the second form depends on the number of elements in the list indexList: if the latter contains exactly one index then the return value is the formatted tablelist item indicated by that index (or an empty string if the index refers to a non-existent item); otherwise the command returns the list of all of the formatted tablelist items corresponding to the indices specified by indexList.

pathName getformattedcells firstCell lastCell
pathName getformattedcells cellIndexList

The first form of the command returns a list whose elements are all of the formatted tablelist elements (i.e., formatted cell contents) between firstCell and lastCell, inclusive.  The value returned by the second form depends on the number of elements in the list cellIndexList: if the latter contains exactly one cell index then the return value is the formatted tablelist element indicated by that cell index; otherwise the command returns the list of all of the formatted tablelist elements corresponding to the cell indices specified by cellIndexList.

pathName getformattedcolumns firstColumn lastColumn
pathName getformattedcolumns columnIndexList

The first form of the command returns a list whose elements are lists themselves, where each of the sublists corresponds to exactly one column between firstColumn and lastColumn, inclusive, and consists of all of the formatted tablelist elements contained in that column.  The value returned by the second form depends on the number of elements in the list columnIndexList: if the latter contains exactly one column index then the return value is a list consisting of all of the formatted tablelist elements contained in the column indicated by that column index; otherwise the command returns a list whose elements are lists themselves, where each of the sublists corresponds to exactly one column index in columnIndexList and consists of all of the formatted tablelist elements contained in that column.

pathName getfullkeys first last
pathName getfullkeys indexList

The first form of the command returns a list whose elements are all of the full keys associated with the tablelist items between first and last, inclusive.  The value returned by the second form depends on the number of elements in the list indexList: if the latter contains exactly one index then the return value is the full key associated with the tablelist item indicated by that index (or an empty string if the index refers to a non-existent item); otherwise the command returns the list of all of the full keys associated with the tablelist items corresponding to the indices specified by indexList.

Each item of a tablelist widget has a unique sequence number that remains unchanged until the item is deleted, thus acting as a key that uniquely identifies the item even if the latter's position (i.e., numerical row index) changes.  This command provides read-only access to the full keys obtained by prepending the letter k to these internal item IDs.

pathName getkeys first last
pathName getkeys indexList

The first form of the command returns a list whose elements are all of the sequence numbers associated with the tablelist items between first and last, inclusive.  The value returned by the second form depends on the number of elements in the list indexList: if the latter contains exactly one index then the return value is the sequence number associated with the tablelist item indicated by that index (or an empty string if the index refers to a non-existent item); otherwise the command returns the list of all of the sequence numbers associated with the tablelist items corresponding to the indices specified by indexList.

Each item of a tablelist widget has a unique sequence number that remains unchanged until the item is deleted, thus acting as a key that uniquely identifies the item even if the latter's position (i.e., numerical row index) changes.  This command provides read-only access to these internal item IDs.

pathName hasattrib name

Returns 1 if the attribute name exists and 0 otherwise.

pathName hascellattrib cellIndex name

Returns 1 if the attribute name for the cell given by cellIndex exists and 0 otherwise.

pathName hascolumnattrib columnIndex name

Returns 1 if the attribute name for the column given by columnIndex exists and 0 otherwise.

pathName hasrowattrib index name

Returns 1 if the attribute name for the row given by index exists and 0 otherwise.

pathName hidetargetmark

Hides the horizontal gap or vertical bar displayed by the showtargetmark command.  No error is generated if the target indicator is not visible at the time the command is invoked.

This command is designed to be used during a drag & drop operation for which the tablelist widget (or its body component) was registered as a drop target.  See the DRAG & DROP SUPPORT section for details and examples.

pathName imagelabelpath cellIndex

Returns the path name of the label widget containing the image embedded into the cell given by cellIndex, as specified with the -image option of the cellconfigure subcommand.  If no image is currently embedded into the cell then the return value is an empty string.

REMARK:  This subcommand comes in handy if you want to associate a binding script with the embedded images.  The recommended way to achieve this is shown in the following example:

bind [.tbl bodytag] <Button-1> {printClickedImage %W %x %y}

proc printClickedImage {w x y} {
    foreach {tbl x y} [tablelist::convEventFields $w $x $y] {}
    set cellIdx [$tbl containingcell $x $y]
    scan $cellIdx "%d,%d" row col
    if {$row >= 0 && $col >= 0 &&
        [string compare $w [$tbl imagelabelpath $cellIdx]] == 0} {
        puts "clicked the image of cell $cellIdx"
    }
}

pathName index index

Returns the integer row index value that corresponds to index.  For compatibility with the Tk core listbox, if index is end then the return value is the number of items in the tablelist (not the index of the last item).  To get the numerical index of the last item, you can pass the word last as index argument.

pathName insert index ?item item ...?

Inserts zero or more new items in the widget's internal list just before the item given by index if the tablelist's state is not disabled.  If index equals the number of items or is specified as end then the new items are added to the end of the widget's list.  Tabulator characters are displayed as \t (i.e., a backslash followed by a t) but are inserted unchanged into the internal list.  Newline characters will force line breaks, i.e., will give rise to multi-line elements (which are displayed in embedded message widgets, created on demand).  The return value is the list of full keys associated with the items just inserted.

REMARK:  This subcommand is not suitable for inserting items into a tablelist designed for displaying a tree hierarchy.  For such tablelist widgets use the insertchildren or insertchildlist subcommand.

pathName insertchildlist parentNodeIndex childIndex itemList

Inserts the items of the list itemList in the widget's internal list of children of the node specified by parentNodeIndex just before the item given by childIndex if the tablelist's state is not disabled.  childIndex must be a number, last (specifying the parent's last child), or end; if it equals the number of children of the node given by parentNodeIndex or is specified as end then the new items are added to the end of the parent's list of children.  Tabulator characters are displayed as \t (i.e., a backslash followed by a t) but are inserted unchanged into the internal list.  Newline characters will force line breaks, i.e., will give rise to multi-line elements (which are displayed in embedded message widgets, created on demand).  The return value is the list of full keys associated with the items just inserted.

This command has the same effect as

eval [list pathName insertchildren parentNodeIndex childIndex] itemList

but it is more efficient and easier to use.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this subcommand is not supported for Tk versions earlier than 8.3.

REMARK:  You can achieve a quite significant speadup by using this subcommand to insert a whole list of items rather than using multiple invocations of insertchildren.

pathName insertchildren parentNodeIndex childIndex ?item item ...?

Inserts zero or more new items in the widget's internal list of children of the node specified by parentNodeIndex just before the item given by childIndex if the tablelist's state is not disabled.  childIndex must be a number, last (specifying the parent's last child), or end; if it equals the number of children of the node given by parentNodeIndex or is specified as end then the new items are added to the end of the parent's list of children.  Tabulator characters are displayed as \t (i.e., a backslash followed by a t) but are inserted unchanged into the internal list.  Newline characters will force line breaks, i.e., will give rise to multi-line elements (which are displayed in embedded message widgets, created on demand).  The return value is the list of full keys associated with the items just inserted.

For technical reasons (the use of the -elide text widget tag option for collapsing a row), this subcommand is not supported for Tk versions earlier than 8.3.

REMARK:  It is explicitly allowed to abbreviate the name insertchildren as insertchild.  This comes in handy when using this subcommand to insert just one child item.

pathName insertcolumnlist columnIndex columnList

Inserts the columns specified by the list columnList just before the column given by columnIndex if the tablelist's state is not disabled.  If columnIndex equals the number of columns or is specified as end then the new columns are added to the end of the column list.  The argument columnList must be a list containing the width, title, and optional alignment specifications for the new columns, in the same form as in the case of the -columns configuration option.  The return value is an empty string.  The elements of the new columns are initially empty strings; the easiest way to change these values is to use the fillcolumn subcommand or the -text column configuration option.

This command has the same effect as

eval [list pathName insertcolumns columnIndex] columnList

but it is more efficient and easier to use.

pathName insertcolumns columnIndex ?width title ?alignment? width title ?alignment? ...?

Inserts zero or more new columns just before the column given by columnIndex if the tablelist's state is not disabled.  If columnIndex equals the number of columns or is specified as end then the new columns are added to the end of the column list.  The arguments following the column index have the same meanings as in the case of the -columns configuration option.  The return value is an empty string.  The elements of the new columns are initially empty strings; the easiest way to change these values is to use the fillcolumn subcommand or the -text column configuration option.

pathName insertlist index itemList

Inserts the items of the list itemList in the widget's internal list just before the item given by index if the tablelist's state is not disabled.  If index equals the number of items or is specified as end then the new items are added to the end of the widget's list.  Tabulator characters are displayed as \t (i.e., a backslash followed by a t) but are inserted unchanged into the internal list.  Newline characters will force line breaks, i.e., will give rise to multi-line elements (which are displayed in embedded message widgets, created on demand).  The return value is the list of full keys associated with the items just inserted.

This command has the same effect as

eval [list pathName insert index] itemList

but it is more efficient and easier to use.

REMARK 1:  You can achieve a quite significant speadup by using this subcommand to insert a whole list of items rather than using multiple invocations of insert.

REMARK 2:  This subcommand is not suitable for inserting items into a tablelist designed for displaying a tree hierarchy.  For such tablelist widgets use the insertchildlist or insertchildren subcommand.

pathName iselemsnipped cellIndex fullTextName

Returns the value 1 if the text displayed in the cell specified by cellIndex is snipped and 0 otherwise.  In both cases, the full (unsnipped) cell text is stored in the variable having the name given by fullTextName; this full text can be the cell's contents or the string obtained from the latter by using the -formatcommand option of the cell's column.  The most common invocation of this command occurs within the procedure specified as the value of the -tooltipaddcommand configuration option.

pathName isexpanded index

Returns the value 1 if the row indicated by index is expanded and 0 otherwise.

pathName istitlesnipped columnIndex fullTextName

Returns the value 1 if the text displayed in the header label specified by columnIndex is snipped and 0 otherwise.  In both cases, the full (unsnipped) label text is stored in the variable having the name given by fullTextName.  The most common invocation of this command occurs within the procedure specified as the value of the -tooltipaddcommand configuration option.

pathName isviewable index

Returns the value 1 if the row indicated by index is viewable and 0 otherwise.  A tablelist row is called viewable if the value of its -hide option is false and all its ancestors are (partly) expanded.   Likewise, a tablelist cell is called viewable if its row is viewable and the value of its column's -hide option is false.

pathName itemlistvar

Returns the name of a variable used by Tablelist to hold the widget's internal list.  The recommended way to use this variable is to create a link to it with the aid of the upvar command, like in the following example:

upvar #0 [.tbl itemlistvar] itemList

In this example, the value of the variable itemList will be the internal list of the tablelist widget .tbl.  Each element of the widget's internal list corresponds to one item, and it is in turn a list whose elements correspond to the elements of that item, except that it has one additional element, holding the item's full key.

REMARK:  The itemlistvar command provides an efficient way for accessing this internal list, instead of retrieving the items with the get subcommand or using the -listvariable option (these methods consume significantly more memory).  It can be useful in situations where the elements of a tablelist widget are to be accessed for creating text files, HTML output, XML data, database commands, etc.  This should, however, be a strictly readonly access; otherwise the results will be unpredictable!

pathName labelpath columnIndex

Returns the path name of the header label corresponding to the column indicated by columnIndex.

pathName labels

Returns a list containing the path names of all header labels of the widget.

pathName labeltag

Returns the name of a binding tag whose name depends on the path name of the tablelist widget and which is associated with the header labels as well as with the additional widgets placed by Tablelist into the latters for displaying header images and sort arrows.  This binding tag is designed to be used when defining non-default binding scripts for the header labels.  From within such event handling scripts you can retrieve the column number and the tablelist widget's path name with the aid of the helper commands tablelist::getTablelistColumn and tablelist::getTablelistPath.

pathName move sourceIndex targetIndex
pathName move sourceIndex targetParentNodeIndex targetChildIndex

The first form of the command moves the item indicated by sourceIndex just before the one given by targetIndex if the tablelist's state is not disabled.  If targetIndex equals the nunber of items or is specified as end then the source item is moved after the last one.  The item specified by targetIndex must have the same parent as the one given by sourceIndex, or else it must be the item just below the last descendant of the source node's parent.

The command's second form moves the item indicated by sourceIndex just before the node having the parent indicated by targetParentNodeIndex and the index targetChildIndex in the parent's list of children if the tablelist's state is not disabled.  targetChildIndex must be a number, last (specifying the target parent's last child), or end; if it equals the number of children of the target parent node or is specified as end then the source item is moved after the target parent node's last child.

Both forms of the command preserve the node hierarchy under the source item, by moving its descendants accordingly.  The return value is an empty string.

pathName movecolumn sourceColumn targetColumn

Moves the column indicated by sourceColumn just before the one given by targetColumn if the tablelist's state is not disabled.  If targetColumn equals the number of columns or is specified as end then the source column is moved after the last one.  Returns an empty string.

pathName nearest y

Given a y-coordinate within the tablelist window, this command returns the index of the viewable tablelist item nearest to that y-coordinate.  The coordinate y is expected to be relative to the tablelist window itself (not its body component).

pathName nearestcell x y

Given an x- and a y-coordinate within the tablelist window, this command returns the index of the viewable tablelist cell nearest to the point having these coordinates.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).

pathName nearestcolumn x

Given an x-coordinate within the tablelist window, this command returns the index of the non-hidden tablelist column nearest to that x-coordinate.  The coordinate x is expected to be relative to the tablelist window itself (not its body component).

pathName noderow parentNodeIndex childIndex

Returns the numerical row index of the node having the parent indicated by parentNodeIndex and the index childIndex in the parent's list of children.  childIndex must be a number, last (specifying the parent's last child), or end; if it equals the number of children of the parent node or is specified as end then the return value will be the row index of the item following the parent node's last descendant.

pathName parentkey nodeIndex

Returns the full key of the parent of the tree node indicated by nodeIndex.  If this argument is specified as root then the return value will be an empty string.  If nodeIndex identifies a top-level item then the subcommand will return root.  For all other items the return value will be a full key of the form knumber.

pathName refreshsorting ?parentNodeIndex?

Sorts the children of the tablelist node specified by parentNodeIndex according to the parameters of the most recent sort, sortbycolumn, or sortbycolumnlist invocation.  If the items haven't been sorted at all, or the sort information was reset by invoking resetsortinfo, then no sorting takes place.  The optional argument parentNodeIndex defaults to root, meaning that all the items are to be sorted per default.  The return value is an empty string.

pathName rejectinput

If invoked from within the Tcl command specified by the -editendcommand configuration option, then this subcommand prevents the termination of the interactive editing of the contents of the cell whose index was passed to the editcell subcommand.  It invokes the seecell subcommand to make sure the respective cell becomes visible (in case it was scrolled out of view), and sets the focus to the temporary widget embedded into the cell.  This command enables you to reject the widget's text during the final validation of the string intended to become the new cell contents.  The return value is an empty string.

pathName resetsortinfo

Resets the information about the sorting of the items.  Subsequent invocations of sortcolumn and sortorder will return -1 and an empty string, respectively.  Similarly, subsequent invocations of sortcolumnlist and sortorderlist will return an empty string.  This command also removes any existing up- or down-arrows displayed by an earlier invocation of sortbycolumn or sortbycolumnlist.  The return value is an empty string.

pathName rowattrib index ?name? ?value name value ...?

Queries or modifies the attributes of the row given by index.  If no name is specified, the command returns a list of pairs, each of which contains the name and the value of an attribute for the row.  If name is specified with no value, then the command returns the value of the one named row attribute, or an empty string if no corresponding value exists (you can use the hasrowattrib subcommand to distinguish this case from the one that the value of an existing row attribute is an empty string).  If one or more name-value pairs are specified, then the command sets the given row attribute(s) to the given value(s); in this case the return value is an empty string.  name may be an arbitrary string.

pathName rowcget index option

Returns the current value of the row configuration option given by option for the row specified by index.  option may have any of the values accepted by the rowconfigure command.

pathName rowconfigure index ?option? ?value option value ...?

Queries or modifies the configuration options of the row given by index.  If no option is specified, the command returns a list describing all of the available options for the row (see Tk_ConfigureInfo for information on the format of this list).  If option is specified with no value, then the command returns a list describing the one named option (this list will be identical to the corresponding sublist of the value returned if no option is specified).  If one or more option-value pairs are specified, then the command modifies the given row option(s) to have the given value(s); in this case the return value is an empty string.  option may have any of the values described in the ROW CONFIGURATION OPTIONS section.

pathName scan option args

This command is used to implement scanning on tablelist widgets.  It has two forms, depending on option:

pathName scan mark x y

Records x and y and the current view in the tablelist window; used in conjunction with later  scan dragto  commands.  Typically this command is associated with a mouse button press in the body component of the widget.  It returns an empty string.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).

pathName scan dragto x y

This command computes the difference between its x and y arguments to the last  scan mark  command for the widget.  It then adjusts the view (the vertical one only in the body component) by 10 times the difference in coordinates.  This command is typically associated with mouse motion events in the body component of the widget, to produce the effect of dragging the table at high speed through the window.  The return value is an empty string.  The coordinates x and y are expected to be relative to the tablelist window itself (not its body component).

pathName searchcolumn columnIndex pattern ?options?

This subcommand searches the elements of the column given by columnIndex to see if one of them matches pattern.  If a match is found, the row index of the first matching element is returned as result (unless the option -all is specified).  If not, the return value is -1.  One or more of the following options may be specified to control the search:

-all	Changes the result to be the list of all matching row indices, which will be in numeric order (or in reverse numeric order when used with the -backwards option).
-backwards	The search will proceed backward through the given column's elements.
-check command	Specifies an additional condition to be fulfilled by the matching elements.  If an element of the specified column matches the given pattern and command is a nonempty string, then the latter is automatically concatenated with the name of the tablelist widget, the element's row index, the numerical equivalent of columnIndex, as well as the element itself or its formatted version (depending on the presence of the -formatted option), the resulting script is evaluated in the global scope, and the return value (which must be a boolean) will determine whether the element in question will still be viewed as matching or not.  The default command is an empty string.  This option enables you to pass arbitrary additional matching criteria to the searching process.
-descend	Search the elements of the specified column in all descendants of the tree node given by the -parent option.  The elements will be visited in the order of their row indices (or in reverse order of their row indices when used with the -backwards option).  The default is to restrict the search to the parent node's children only.
-exact	The matching element(s) must be identical to the literal string pattern.
-formatted	Examine the formatted versions of the elements rather than the internal cell values.
-glob	Treat pattern as a glob-style pattern and match it against the elements using the same rules as the  string match  command.
-nocase	Causes comparisons to be handled in a case-insensitive manner.  Has no effect if combined with the -numeric option.
-not	This option negates the sense of the match, returning the row index of the first non-matching element (or, in the presence of the -all option, the list of row indices of all non-matching elements) of the given column.
-numeric	The elements are to be compared to pattern as integer or floating-point values, using the == comparison operator.  This option is only meaningful when used with -exact.
-parent nodeIndex	This option restricts the search to the children (or descendants, when used with -descend) of the tree node given by nodeIndex.  The default parent is root.
-regexp	Treat pattern as a regular expression and match it against the elements using the rules described in the re_syntax reference page.
-start index	The elements of the specified column are to be searched (forwards or backwards) starting at the row given by index.  This option makes it easy to provide incremental search.

If all matching style options -exact, -glob, and -regexp are omitted then the matching style defaults to -glob.  If more than one of them is specified, the last matching style given takes precedence.

Before examining the children (or descendants, when used with the -descend option) of a row whose children have not been inserted yet, the command specified as the value of the -populatecommand option (if any) is automatically concatenated with the name of the tablelist widget and the row index, and the resulting script is evaluated in the global scope.  This enables you to insert the children on demand, just before searching them for the specified pattern.

pathName see index

Adjusts the view in the tablelist so that the item given by index is visible.  If the item is already visible then the command has no effect; if the item is near one edge of the window then the tablelist scrolls to bring the item into view at the edge; otherwise the tablelist scrolls to center the item.

pathName seecell cellIndex

Adjusts the view in the tablelist so that the cell given by cellIndex is visible.  If the cell is already visible then the command has no effect; if the cell is near one edge of the window then the tablelist scrolls to bring the cell into view at the edge; otherwise the tablelist scrolls to center the cell.  If the value of the -titlecolumns option is positive then the centering of the cell is only done vertically; the horizontal scrolling (which in this case is performed column-wise) will just bring the cell into view next to the title columns or at the right edge of the window.

id="seecolumn">pathName seecolumn columnIndex

Adjusts the view in the tablelist so that the column given by columnIndex is visible.  If the column is already visible then the command has no effect; if the column is near one edge of the window then the tablelist scrolls horizontally to bring the column into view at the edge; otherwise the tablelist scrolls horizontally to center the column.  If the value of the -titlecolumns option is positive then the horizontal scrolling (which in this case is performed column-wise) will just bring the column into view next to the title columns or at the right edge of the window.

pathName selection option args

This command is used to adjust the selection within a tablelist widget.  It has several forms, depending on option:

pathName selection anchor index

Sets the selection anchor to the item given by index.  If index refers to a nonexistent or non-viewable item, then the closest viewable item is used.  The selection anchor is the end of the selection that is fixed while dragging out a selection with the mouse if the selection type is row.  The index anchor may be used to refer to the anchor item.

pathName selection clear first last
pathName selection clear indexList

If any of the items between first and last (inclusive) or corresponding to the indices specified by the list indexList contain at least one selected cell, they are deselected.  The selection state is not changed for items outside the range given in the first form of the command or different from those specified by the index list given in its second form.

pathName selection includes index

Returns 1 if the item indicated by index contains at least one selected cell, 0 if it doesn't.

pathName selection set first last
pathName selection set indexList

Selects all of the selectable items in the range between first and last, inclusive, or corresponding to the indices specified by the list indexList, without affecting the selection state of any other items.

If the tablelist's state is disabled and option is different from includes then the command just returns an empty string, without performing any of the above actions.

pathName separatorpath ?columnIndex?

If the optional argument is not specified, then this command returns the path name of the special separator displayed to mark the end of the title columns if the value of the -titlecolumns option is positive and an empty string otherwise.  If the optional argument is present, then the command returns the path name of the separator attached to the right edge of the header label indicated by columnIndex if the value of the -showseparators configuration option is true and an empty string otherwise.

pathName separators

Returns a list containing the path names of all column separators.  If the value of the -titlecolumns option is positive then the first element of the list will be the path name of the special separator displayed to mark the end of the title columns.  Whether the path names of the other separators are included in the list, depends on the value of the -showseparators configuration option.

pathName showtargetmark before|inside index

Displays a drop target indicator having the form of a horizontal gap or vertical bar before or inside the row specified by index.  If the subcommand's name is followed by before and index equals the number of items or is specified as end then the horizontal gap will be shown just after the tablelist's last row.  If the subcommand is used with the inside option then the index end is interpreted as indicating the widget's last item.

This command is designed to be used during a drag & drop operation for which the tablelist widget (or its body component) was registered as a drop target.  See the DRAG & DROP SUPPORT section for details and examples.

pathName size

Returns the total number of items in the tablelist widget.

pathName sort ?-increasing|-decreasing?

Sorts the items in increasing or decreasing order, as specified by the optional argument.  The default is -increasing.  Uses the value of the -sortcommand widget configuration option as comparison command.  sort also removes any existing up- or down-arrows displayed by an earlier invocation of sortbycolumn or sortbycolumnlist.  After sorting the items, the command conditionally adjusts the vertical view as follows: (a) if interactive cell editing is in progress then the cell being edited is brought into view; (b) else, if exactly one item is selected then the view is shifted to bring that item into view; (c) else, if the tablelist's body is the most recent window to have the input focus among all the windows in the same top-level as the widget itself then the currently active item is brought into view.

pathName sortbycolumn columnIndex ?-increasing|-decreasing?

Sorts the items based on the elements of the column given by columnIndex, in increasing or decreasing order, as specified by the optional argument.  The default is -increasing.  The sorting process is controlled by the values of the -sortmode and -sortcommand options for the given column.  If both the value of the -showarrow configuration option and that of the -showarrow option for the specified column are true then an up- or down-arrow indicating the sort order will be placed into the column's label.  The shape of the arrow depends on the command's optional argument and on the value of the -incrarrowtype configuration option.  If the label's text is right-aligned then the arrow will be displayed on the left side of the label, otherwise on its right side.  After sorting the items, the vertical view is adjusted in the same way as in the case of the sort subcommand.

The actions described above are only performed if the specified column's -showlinenumbers option hasn't been set to true.

pathName sortbycolumnlist columnIndexList ?sortOrderList?

Sorts the items based on the elements of the columns given by the columnIndexList argument, which must be a list of distinct column indices.  Only those elements of this list are considered significant that identify columns whose -showlinenumbers option hasn't been set to true.

The items are first sorted based on the column specified by the last significant element of columnIndexList, then based on the one given by the last but one significant element, and so on.  The order of each sort operation is taken from the optional argument sortOrderList, whose elements must be (abbreviations of) increasing or decreasing.  If this argument was not specified or contains less elements than columnIndexList then the missing sort orders are assumed to be increasing.  Each sorting process is controlled by the values of the -sortmode and -sortcommand options for the respective column.  If the column's index was specified among the first 9 significant elements of columnIndexList and both the value of the -showarrow configuration option and that of the -showarrow option for that column are true then an up- or down-arrow indicating the sort order will be placed into the column's label.  The shape of the arrow depends on the respective sort order and on the value of the -incrarrowtype configuration option.  If the label's text is right-aligned then the arrow will be displayed on the left side of the label, otherwise on its right side.  If more than one sort arrows are to be displayed then the first 9 sort ranks (1 for the first significant element of columnIndexList, 2 for the second one, and so on) will also be shown to the right of the arrows.  After sorting the items, the vertical view is adjusted in the same way as in the case of the sort subcommand.

pathName sortcolumn

Returns the numerical index of the column by which the items were last sorted with the aid of the sortbycolumn or sortbycolumnlist command, or -1 if they were last sorted with the sort command or haven't been sorted at all, or the sort information was reset by invoking resetsortinfo.  If called from within the command specified as the value of the -sortcommand widget or column configuration option, then the return value of this subcommand refers to the sorting in progress rather than the most recent one.

pathName sortcolumnlist

Returns a list consisting of the numerical indices of the columns by which the items were last sorted with the aid of the sortbycolumnlist or sortbycolumn command (in the second case the list will contain exactly one element), or an empty string if they were last sorted with the sort command or haven't been sorted at all, or the sort information was reset by invoking resetsortinfo.  If called from within the command specified as the value of the -sortcommand widget or column configuration option, then the return value of this subcommand refers to the sorting in progress rather than the most recent one.

pathName sortorder

Returns the sort order (as increasing or decreasing) from the last sorting performed by the sort, sortbycolumn, or sortbycolumnlist command, or an empty string if the items haven't been sorted at all, or the sort information was reset by invoking resetsortinfo.  If called from within the command specified as the value of the -sortcommand widget or column configuration option, then the return value of this subcommand refers to the sorting in progress rather than the most recent one.

pathName sortorderlist

Returns a list consisting of the sort orders (as increasing or decreasing) from the last invocation of the sortbycolumnlist or sortbycolumn command (in the second case the list will contain exactly one element), or an empty string if the items were last sorted with the sort command or haven't been sorted at all, or the sort information was reset by invoking resetsortinfo.  If called from within the command specified as the value of the -sortcommand widget or column configuration option, then the return value of this subcommand refers to the sorting in progress rather than the most recent one.

pathName targetmarkpath

Returns the path name of the drop target indicator (displayed as a horizontal gap or vertical bar) belonging to the tablelist's body.

pathName targetmarkpos y ?-any|-horizontal|-vertical?

Given a y-coordinate within the tablelist window, this command returns a two-element list containing the arguments to be passed to the showtargetmark subcommand in order to display the drop target indicator corresponding to that y-coordinate.  The first list element will be the string before or inside, and the second one a numerical row index.  These list elements depend on the relative position of y within the containing row (if any), as well as on the optional second argument:  The default option -any allows both before and inside as first element of the result list, while the options -horizontal and -vertical restrict the value of the first list element to before and inside, respectively.  If the option -vertical was specified and no tablelist item contains the given y-position, then the return value is the list  {inside -1}.  The coordinate y is expected to be relative to the tablelist window itself (not its body component).

This command is designed to be used during a drag & drop operation for which the tablelist widget (or its body component) was registered as a drop target.  See the DRAG & DROP SUPPORT section for details and examples.

pathName togglecolumnhide firstColumn lastColumn
pathName togglecolumnhide columnIndexList

Toggles the value of the -hide option for one or more columns of the tablelist widget.  In the first form of the command, firstColumn and lastColumn are indices specifying the first and last columns in the range whose visibility is to be toggled.  The command's second form accepts a list columnIndexList of indices specifying the columns whose visibility is to be toggled.  Returns an empty string.  After toggling the hidden state of the specified columns, the <<TablelistColHiddenStateChanged>> virtual event is generated.  For Tk versions 8.5 or higher, this virtual event is generated with its -data option set to a list consisting of the numerical column indices of the columns whose -hide option was toggled.  The main advantage of using this command instead of invoking columnconfigure for each of the specified columns is that it causes only one redisplay of the widget's contents, thus being significantly faster.

pathName togglerowhide first last
pathName togglerowhide indexList

Toggles the value of the -hide option for one or more rows of the tablelist widget.  In the first form of the command, first and last are indices specifying the first and last rows in the range whose visibility is to be toggled.  The command's second form accepts a list indexList of indices specifying the rows whose visibility is to be toggled.  Returns an empty string.  After toggling the hidden state of the specified rows, the <<TablelistRowHiddenStateChanged>> virtual event is generated.  For Tk versions 8.5 or higher, this virtual event is generated with its -data option set to a list consisting of the numerical row indices of the rows whose -hide option was toggled.  Just like the -hide row configuration option, this subcommand is not supported for Tk versions earlier than 8.3.

CAUTION:  Tk versions 8.3 - 8.4.12 had a bug that caused a segmentation fault if the whole content of a text widget was elided.  This bug was also present in Tk 8.5.a1 - 8.5.a3.  When using one of these earlier Tk versions, this bug will produce a crash if all the rows of a tablelist widget are hidden.  It is your responsibility to avoid such situations when using a Tk version having this bug!

pathName toplevelkey index

If the item identified by index is a top-level one then the subcommand returns the full key of that item.  Otherwise the return value is the full key of the unique top-level item having the given item among its descendants.

pathName unsetattrib name

Unsets the attribute name.  Returns an empty string.

pathName unsetcellattrib cellIndex name

Unsets the attribute name for the cell given by cellIndex.  Returns an empty string.

pathName unsetcolumnattrib columnIndex name

Unsets the attribute name for the column given by columnIndex.  Returns an empty string.

pathName unsetrowattrib index name

Unsets the attribute name for the row given by index.  Returns an empty string.

pathName viewablerowcount ?first last?

Returns the number of viewable rows in the index range given by first and last.  If these optional indices are not specified then first defaults to 0 and last defaults to one less the total number of items, i.e., the index range comprises all rows.

pathName windowpath cellIndex

Returns the path name of the window contained in the cell given by cellIndex, created with the -window option of the cellconfigure subcommand.  If no window is currently embedded into the cell then the return value is an empty string.

pathName xview args

This command is used to query and change the horizontal position of the information in the widget's window.  It can take any of the following forms:

pathName xview

Returns a list containing two elements.  Each element is a real fraction between 0 and 1; together they describe the horizontal span that is visible in the window.  For example, if the first element is .2 and the second element is .6, 20% of the tablelist's scrollable text is off-screen to the left, the middle 40% is visible in the window, and 40% of the scrollable text is off-screen to the right.  These are the same values passed to scrollbars via the -xscrollcommand option.

pathName xview units

If the value of the -titlecolumns option is positive then this command adjusts the view in the window so that the column whose offset from the end of the title column area equals units non-hidden columns is displayed next to the title columns.  Otherwise the command adjusts the view in the window so that the character position given by units is displayed at the left edge of the window.  Character positions are defined by the width of the character 0.

pathName xview moveto fraction

Adjusts the view in the window so that fraction of the total width of the scrollable tablelist text is off-screen to the left.  fraction must be a fraction between 0 and 1.

pathName xview scroll number what

This command shifts the view in the window left or right according to number and what.  number must be an integer.  what must be either units or pages or an abbreviation of one of these.  If what is units, the view adjusts left or right by number non-hidden columns or character units (the width of the 0 character) on the display, depending on the value of the -titlecolumns option; if what is pages then the view adjusts by number screenfuls.  If number is negative then columns or characters farther to the left become visible; if it is positive then columns or characters farther to the right become visible.

pathName yview args

This command is used to query and change the vertical position of the text in the window of the widget's body component.  It can take any of the following forms:

pathName yview

Returns a list containing two elements, both of which are real fractions between 0 and 1.  The first element gives the position of the viewable tablelist item at the top of the window, relative to the tablelist as a whole (0.5 means it is halfway through the tablelist, for example).  The second element gives the position of the viewable tablelist item just after the last one in the window, relative to the tablelist as a whole.  These are the same values passed to scrollbars via the -yscrollcommand option.

pathName yview units

Adjusts the view in the window so that the item whose offset equals units viewable rows is displayed at the top of the window.

pathName yview moveto fraction

Adjusts the view in the window so that the viewable item given by fraction appears at the top of the window.  fraction is a fraction between 0 and 1; 0 indicates the first viewable item in the tablelist, 0.33 indicates the viewable item one-third the way through the tablelist, and so on.

pathName yview scroll number what

This command shifts the view in the window up or down according to number and what.  number must be an integer.  what must be either units or pages or an abbreviation of one of these.  If what is units, the view adjusts up or down by number viewable rows; if it is pages then the view adjusts by number screenfuls.  If number is negative then earlier items become visible; if it is positive then later items become visible.

DEFAULT AND INDIVIDUAL BINDINGS FOR THE TABLELIST BODY

The body component of a tablelist is implemented as a text widget whose binding tag Text is replaced with a new binding tag called TablelistBody.  The latter has all the events of the Listbox widget class, and several of its binding scripts are obtained from those of Listbox by replacing the event fields %W, %x, and %y with the path name of the tablelist widget and the x and y coordinates relative to the latter.  These values are assigned to the help variables tablelist::W, tablelist::x, and tablelist::y by invoking the helper command tablelist::convEventFields as follows:

foreach {tablelist::W tablelist::x tablelist::y} \
    [tablelist::convEventFields %W %x %y] {}

This conversion of the event fields is necessary because the Tcl command associated with a tablelist expects any coordinates relative to the widget itself, not its body component.  It makes use of help variables from the tablelist namespace in order to avoid any conflicts with global variables.

Several of the events have no %x and %y fields; in this case another helper command tablelist::getTablelistPath is used to set the help variable tablelist::W to the path name of the tablelist widget:

set tablelist::W [tablelist::getTablelistPath %W]

The binding tag TablelistBody replaces the class name (Frame or TSeparator) of the separator widgets, too.  It also replaces the binding tag Message of the message widgets used to display multi-line elements, as well as the binding tag Label of the label widgets used to display embedded images.  This makes sure that the default handling of the mouse events on the column separators, multi-line cells, and embedded images is the same as in the rest of the tablelist's body.

When defining individual bindings for tablelist widgets, the same conversion of the event fields is needed as for the default bindings.  For example, the binding script below for the tablelist widget .tbl prints the index of the cell where mouse button 1 was clicked:

bind [.tbl bodytag] <Button-1> {
    foreach {tablelist::W tablelist::x tablelist::y} \
        [tablelist::convEventFields %W %x %y] {}
    puts "clicked on cell [.tbl containingcell $tablelist::x $tablelist::y]"
}

By associating the script with the binding tag returned by the bodytag subcommand instead of just with the path name of the tablelist's body we make sure to have the same event handling for the separators, multi-line cells, and embedded images as for the rest of the tablelist's body.

The following improved version of the binding script above uses a procedure and thus eliminates the need for prefixing the converted variables with "tablelist::":

bind [.tbl bodytag] <Button-1> {printClickedCell %W %x %y}

proc printClickedCell {w x y} {
    foreach {tbl x y} [tablelist::convEventFields $w $x $y] {}
    puts "clicked on cell [$tbl containingcell $x $y]"
}

The bindings associated with the binding tag TablelistBody, created automatically by the tablelist::tablelist command, ensure that the body component of a tablelist has the same default behavior as a listbox widget.  If the selection type is row (which is the default) then everything described in the "DEFAULT BINDINGS" section of the listbox manual entry applies to the body component of a tablelist, too.  The only difference is that the word "element" in that manual page has to be replaced with "item" when applying the description to the body of a tablelist widget.

If the selection type is cell then everything described in the "DEFAULT BINDINGS" section of the listbox manual entry applies to the body component of a tablelist, too, with the following extensions and changes:

1. If Tab or Shift-Tab is pressed, the location cursor (active element) moves to the next/previous element.  If the selection mode is browse or extended then the new active element is also selected and all other elements are deselected.  In extended mode the new active element becomes the selection anchor.  Notice that these bindings replace the common inter-widget navigation via Tab and Shift-Tab with inter-cell navigation.  Just like in the case of the text widget, Control-Tab and Control-Shift-Tab are intended to be used for widget-to-widget keyboard navigation.  Unfortunately, this won't always work because some window managers intercept the latter key sequences and use them for their own purposes (like inter-workplace navigation).  For this reason, Tablelist supports the additional key sequences Meta-Tab and Meta-Shift-Tab as replacements for Control-Tab and Control-Shift-Tab, respectively.
2. If the Left or Right key is pressed, the location cursor (active element) moves to the previous/next element of the active row.  If the selection mode is browse or extended then the new active element is also selected and all other elements are deselected.  In extended mode the new active element becomes the selection anchor.
3. In extended mode, Shift-Left and Shift-Right move the location cursor (active element) to the previous/next element of the active row and also extend the selection to that element in a fashion similar to dragging with mouse button 1.
4. If the Home or End key is pressed, the location cursor (active element) moves to the first/last element of the active row, the new active element is selected, and all other elements are deselected.
5. In extended mode, Shift-Home and Shift-End extend the selection to the first/last element of the active row.  In multiple mode, Shift-Home and Shift-End move the location cursor to the first/last element of the active row.
6. If the location cursor is in an editable cell then Return and KP_Enter start the interactive editing of the active element.

Just like in the case of the listbox widget, any changes to the selection will automatically generate the virtual event <<ListboxSelect>>.  Instead of this event (which is supported for compatibility reasons), the virtual event <<TablelistSelect>> can be used to be made aware of any changes to tablelist selection.  Both events will be generated independently of the selection type.

LOCAL DRAG & DROP:  The following binding associated with the binding tag TablelistBody is only valid if the selection mode is single or multiple:

If mouse button 1 is clicked on an item and then dragged outside that item, and the value of the -movablerows configuration option is true, then the mouse cursor takes on the shape specified by the -movecursor option, indicating that the item in question is being moved to another position.  The new item position (if any) is visualized with the aid of a gap placed before the target row or a bar placed inside the latter (depending on the current mouse position), indicating whether the source item would be moved before this row or become a child of it.  This local drag & drop operation ends when mouse button 1 is released, and can be canceled by pressing the Escape key.  In both cases, the mouse cursor is reset to its original value, specified by the -cursor configuration option.  After releasing mouse button 1 in the presence of a valid target, the source item is moved before the target row or just before the latter's first child, and the virtual event <<TablelistRowMoved>> is generated.  For Tk versions 8.5 or higher, this virtual event is generated with its -data option set to a list of length 3, whose elements are derived from the arguments passed to the second form of the move subcommand invoked for moving the source row to its new position.  The first list element will be the full key corresponding to the first argument, the second one will be root or the full key corresponding to the second argument, and the third list element will be identical to the third argument passed to the move subcommand.

Notice that, depending on the current mouse position during the local drag & drop, there can be a corresponding potential target row or not.  For instance, a tree item cannot become a sibling of one of its descendants, and not all items might be allowed to have children or to become top-level ones (example: in a file manager, regular file items cannot be parents of other items and should not be allowed to become top-level ones).  To decide whether the row corresponding to the y-coordinate of the current mouse position represents a valid potential target, the Tablelist code first checks whether moving the source item before that row or making it a child of the latter is allowed from the point of view of the general tree structure.  If this is the case and the move operation would change the source item's parent (and the Tk version is at least 8.3), and the command specified as the value of the -acceptchildcommand configuration option is a nonempty string, then this command is concatenated with the name of the tablelist widget, the node index of the would-be new parent node, and the row index of the dragged item, the resulting script is evaluated in the global scope, and if the return value (which must be a boolean) is false, then the source item will not be allowed to be moved to the current mouse position.  Likewise, if the command specified as the value of the -acceptdropcommand configuration option is a nonempty string, then this command is concatenated with the name of the tablelist widget, the row index of the would-be new target row, and the row index of the dragged item, the resulting script is evaluated in the global scope, and if the return value (which must be a boolean) is false, then the source item will not be allowed to be moved to the current mouse position.

Recall that if the selection mode is multiple then pressing mouse button 1 on a selected item or element normally deselects that item or element (depending on the selection type).  However, if in addition the value of the -movablerows configuration option is true then the clicked row is a potential drag source for the local drag & drop operation described above, and for this reason the clicked item or element will only be deselected when releasing mouse button 1 over the same item or element.

DRAG SOURCE SUPPORT FOR GLOBAL DRAG & DROP:  Besides the local drag & drop, the default bindings also support the TkDND compiled extension and the drag & drop framework included in BWidget, as well as custom drag & drop implementations.  A tablelist widget is viewed as a drag source for mouse button 1 if its body component was registered as such via the  tkdnd::drag_source register  or the Widget DragSite::register command, or the tablelist's -customdragsource option was set to true.  The default bindings provide drag source support as described below:

1. If the selection mode is extended then pressing mouse button 1 on a selected item or element normally deselects all the other items or elements (depending on the selection type).  However, if the tablelist is a drag source for mouse button 1, then the other items or elements will only be deselected when releasing mouse button 1 over the clicked item or element.  This is because the mouse click might be followed by a drag, intended for all currently selected items or elements.
2. Similarly, if the selection mode is multiple then pressing mouse button 1 on a selected item or element normally deselects that item or element (depending on the selection type).  However, if the tablelist is a drag source for mouse button 1, then the clicked item or element will only be deselected when releasing mouse button 1 over the same item or element.  Again, this is because the mouse click might be followed by a drag, intended for all currently selected items or elements.
3. Whenever the mouse leaves the tablelist window with button 1 down. the default bindings normally perform an automatic scrolling, just like in the case of the Tk listbox widget.  However, if the tablelist is a drag source for mouse button 1, then the automatic scrolling will be suppressed, in order to avoid any conflicts with the drag operation.

TREE WIDGET BINDINGS:  The following bindings associated with the binding tag TablelistBody apply to tablelists used as tree widgets:

1. Pressing mouse button 1 over an expand/collapse control toggles the expanded/collapsed state of the corresponding row.
2. If the current active row contains an expand/collapse control in collapsed state then the Right, plus, and KP_Add keys expand the corresponding row by invoking the non-recursive version of the expand subcommand.
3. If the current active row contains an expand/collapse control in expanded state then the Left, minus, and KP_Subtract keys collapse the corresponding row by invoking the non-recursive version of the collapse subcommand.

DEFAULT AND INDIVIDUAL BINDINGS FOR THE HEADER LABELS

The tablelist::tablelist command automatically creates the following bindings for the header labels:

1. If the mouse pointer is on the right edge of a header label or within a few pixels of its right edge, and both the value of the -resizablecolumns configuration option and that of the -resizable option for the column corresponding to that label are true, then the mouse cursor takes on the shape specified by the -resizecursor option.  By clicking mouse button 1 in this area and moving the mouse while its button 1 is down, the column corresponding to that label will be resized by the amount of the cursor motion.  The interactive column resizing ends when mouse button 1 is released, and can be canceled by pressing the Escape key.  In both cases, the mouse cursor is reset to its original value, specified by the -cursor configuration option.  When the column resize operation is finished, the virtual event <<TablelistColumnResized>> is generated, with its -data option set to the numerical column index for Tk versions 8.5 or higher.
2. If mouse button 1 is pressed over a header label but outside the resize area described above and then dragged outside the label, and the value of the -movablecolumns configuration option is true, then the mouse cursor takes on the shape specified by the -movecolumncursor option, indicating that the column in question is being moved to another position, visualized with the aid of a gap placed before the label of the target column.  This operation ends when mouse button 1 is released, and can be canceled by pressing the Escape key when the mouse pointer is outside the label.  In both cases, the mouse cursor is reset to its original value, specified by the -cursor configuration option.  After releasing mouse button 1, the source column is moved before the one indicated by the gap mentioned above and the virtual event <<TablelistColumnMoved>> is generated.  For Tk versions 8.5 or higher, this virtual event is generated with its -data option set to a list of length 4, whose first two elements are identical to the two numerical column indices passed to the movecolumn subcommand invoked for moving the source column to its new position, and the last two elements are the corresponding column names, retrieved with the aid of the  columncget ... -name  subcommand.
3. If mouse button 1 is pressed over a header label but outside the resize area described above and later released over the same label, and the command specified by the -labelcommand option is a nonempty string, then this command is concatenated with the name of the tablelist widget and the column index of the respective label, and the resulting script is evaluated in the global scope.  If another nonempty label command was specified at column level by using the columnconfigure option of the Tcl command associated with the tablelist widget, then that column-specific command will be used instead of the global one.  If mouse button 1 was pressed together with the Shift key then the widget- or column-specific command mentioned above will be replaced with the one specified by the -labelcommand2 option at widget or column level.
4. The Tablelist package defines the virtual event <<Button3>> as <Button-3> for all windowing systems and additionally as <Control-Button-1> for Mac OS Classic and Mac OS X Aqua.  If this event occurs over a header label and both the value of the -resizablecolumns configuration option and that of the -resizable option for the column corresponding to that label are true, then the width of that column is set to zero, i.e., it is made just large enough to hold all the elements in the column, including the header (but no larger than the maximum width indicated by the -maxwidth column configuration option), and the virtual event <<TablelistColumnResized>> is generated, with its -data option set to the numerical column index for Tk versions 8.5 or higher.  The same action is triggered by double-clicking the resize area of a header label.
5. The Tablelist package defines the virtual event <<ShiftButton3>> as <Shift-Button-3> for all windowing systems and additionally as <Shift-Control-Button-1> for Mac OS Classic and Mac OS X Aqua.  If this event occurs over a header label and both the value of the -resizablecolumns configuration option and that of the -resizable option for the column corresponding to that label are true, then the width of that column is set to its last static width (if any) and the virtual event <<TablelistColumnResized>> is generated, with its -data option set to the numerical column index for Tk versions 8.5 or higher.  The same action is triggered by double-clicking the resize area of a header label with the Shift key held down.

If the tablelist's state is disabled then none of the above actions occur: the labels are completely insensitive.

If you want to define non-default bindings for the header labels, it is recommended to associate them with the binding tag whose name is returned by the labeltag subcommand and make use of the helper commands tablelist::getTablelistColumn and tablelist::getTablelistPath.  For example, to replace the default binding for <Button-3> with a script that performs a column-dependent action, you can proceed like in the code shown below:

bind [.tbl labeltag] <Button-3> {
    puts "right-clicked on header label no. [tablelist::getTablelistColumn %W]"
    break
}

DEFAULT BINDINGS FOR INTERACTIVE CELL EDITING

The tablelist::tablelist command extends and partially redefines the bindings of some of the components of the temporary embedded widget used for interactive cell editing, which is started by pressing mouse button 1 in an editable cell (see the -editselectedonly option for details) or using keyboard navigation to move from one editable cell into another.  If the selection type is cell and the location cursor is in an editable cell, then the interactive editing of the active element can also be started by pressing Return or KP_Enter.

The affected components of the temporary embedded widget depend on the edit window: the widget itself in case of a Tk or tile checkbutton or menubutton; the edit window's entry children in case of a mentry widget; the only entry or entry-like component of the edit window in all other cases (see also the entrypath subcommand).  The list of binding tags of these edit window components contains two addditional tags, inserted just before their path names: the binding tag whose name is returned by the editwintag subcommand, followed by the tag TablelistEdit.  The bindings described below are associated with the tag TablelistEdit, and can be overridden for individual tablelist widgets by making use of the binding tag given by the editwintag subcommand.

1. Control-i inserts a tabulator character into the edit window's entry or entry-like components (if any), at the point of the insertion cursor.
2. Control-j inserts a newline character into the edit window's entry or entry-like components (if any), at the point of the insertion cursor.
3. If the edit window is a text or ctext widget then Return and KP_Enter insert a newline character at the point of the insertion cursor.  Otherwise they terminate the editing and destroy the edit window.
4. Control-Return and Control-KP_Enter terminate the editing and destroy the edit window.
5. Escape aborts the editing and destroys the edit window.
6. A click with the left mouse button anywhere in the tablelist's body, outside the cell just being edited, terminates the editing in the current cell and destroys the edit window or moves it into the cell that was just clicked into if the latter is editable.
7. When editing a cell that is not the only editable cell of the tablelist widget, Tab and Shift-Tab terminate the editing in the current cell, move the edit window into the next/previous editable cell of the tablelist, select the contents of the edit window's first entry or entry-like component (if any), and set the insertion cursor to its end.  If the new edit window is a text or ctext widget then its contents are left unselected.  Notice that these bindings replace the common inter-widget navigation via Tab and Shift-Tab with inter-cell navigation.  Just like in the case of the text widget, Control-Tab and Control-Shift-Tab are intended to be used for widget-to-widget keyboard navigation during interactive cell editing.  Unfortunately, this won't always work because some window managers intercept the latter key sequences and use them for their own purposes (like inter-workplace navigation).  For this reason, Tablelist supports the additional key sequences Meta-Tab and Meta-Shift-Tab as replacements for Control-Tab and Control-Shift-Tab, respectively.
8. When editing a cell that is not the first/last editable cell within its row, Alt-Left/Alt-Right (Command-Left/Command-Right on Mac OS Classic and Mac OS X Aqua) terminates the editing in the current cell, moves the edit window into the previous/next editable cell of the row, selects the contents of the edit window's first entry or entry-like component (if any), and sets the insertion cursor to its end.  If the new edit window is a text or ctext widget then its contents are left unselected.  The key sequence Meta-Left/Meta-Right has the same effect as Alt-Left/Alt-Right.  If tk_strictMotif is false and the edit window is not a text or ctext widget then Meta-b and Meta-f behave the same as Alt-Left and Alt-Right, respectively.  If the edit window is a Tk or tile checkbutton or menubutton widget then Left/Right has the same effect as Alt-Left/Alt-Right.
9. When editing a cell that is not the first/last editable cell within its column, Alt-Up/Alt-Down (Command-Up/Command-Down on Mac OS Classic and Mac OS X Aqua) terminates the editing in the current cell, moves the edit window one line up/down within the column, selects the contents of the edit window's first entry or entry-like component (if any), and sets the insertion cursor to its end.  If the new edit window is a text or ctext widget then its contents are left unselected.  The key sequence Meta-Up/Meta-Down has the same effect as Alt-Up/Alt-Down.  If tk_strictMotif is false and the edit window is not a text or ctext widget or an Iwidgets combobox, then Control-p and Control-n behave the same as Alt-Up and Alt-Down, respectively.  If the edit window is a Tk or tile entry, Tk or tile checkbutton, Tk or tile menubutton, BWidget Entry, Iwidgets entryfield/spinner/spinint, or a mentry widget of type "FixedPoint", then Up/Down has the same effect as Alt-Up/Alt-Down.
10. When editing a cell that is not the first/last editable cell within its column, Alt-Prior/Alt-Next (Command-Prior/Command-Next on Mac OS Classic and Mac OS X Aqua) terminates the editing in the current cell, moves the edit window up/down by one page within the column, selects the contents of the edit window's first entry or entry-like component (if any), and sets the insertion cursor to its end.  If the new edit window is a text or ctext widget then its contents are left unselected.  The key sequence Meta-Prior/Meta-Next has the same effect as Alt-Prior/Alt-Next.  If the edit window is not a text or ctext widget, BWidget SpinBox, Oakley combobox, or a mentry widget of type "Date", "Time", "DateTime", "IPAddr", or "IPv6Addr", then Prior/Next has the same effect as Alt-Prior/Alt-Next.
11. When editing a cell that is not the only editable cell of the tablelist widget, Alt-Home/Alt-End (Command-Home/Command-End on Mac OS Classic and Mac OS X Aqua) terminates the editing in the current cell, moves the edit window into the first/last editable cell of the tablelist, selects the contents of the edit window's first entry or entry-like component (if any), and sets the insertion cursor to its end.  If the new edit window is a text or ctext widget then its contents are left unselected.  The key sequence Meta-Home/Meta-End has the same effect as Alt-Home/Alt-End.  If tk_strictMotif is false and the edit window is not a text or ctext widget then Meta-< and Meta-> behave the same as Alt-Home and Alt-End, respectively.  If the edit window is not a text or ctext widget then Control-Home/Control-End has the same effect as Alt-Home/Alt-End.

Before moving the edit window, the key sequences mentioned under 7 - 11 move the active item or element and change the (cell)selection and the (cell)selection anchor in the body of the tablelist widget.  For example, if Alt-Up/Alt-Down or Meta-Up/Meta-Down (Command-Up/Command-Down on Mac OS Classic and Mac OS X Aqua) is pressed when editing a cell that is not the first/last editable cell within its column, then the active item or element (depending on the selection type) moves one line up/down.  If the selection mode is browse or extended then the new active item or element is also selected and all other items or elements are deselected.  In extended mode the new active item or element becomes the (cell)selection anchor.  This is exactly the same behavior as the one exhibited by the Up and Down keys in the tablelist's body.

If the tablelist's state is disabled then none of the above actions occur.

KEYWORDS

tablelist, multi-column, listbox, tree, widget