TableView QML Type

Provides a table view of items to display data from a model. More...

Import Statement: import QtQuick
Inherits:

Flickable

Inherited By:

TreeView

Properties

Attached Properties

Signals

Attached Signals

Methods

Detailed Description

A TableView has a model that defines the data to be displayed, and a delegate that defines how the data should be displayed.

TableView inherits Flickable. This means that while the model can have any number of rows and columns, only a subsection of the table is usually visible inside the viewport. As soon as you flick, new rows and columns enter the viewport, while old ones exit and are removed from the viewport. The rows and columns that move out are reused for building the rows and columns that move into the viewport. As such, the TableView support models of any size without affecting performance.

A TableView displays data from models created from built-in QML types such as ListModel and XmlListModel, which populates the first column only in a TableView. To create models with multiple columns, either use TableModel or a C++ model that inherits QAbstractItemModel.

A TableView does not include headers by default. You can add headers using the HorizontalHeaderView and VerticalHeaderView from Qt Quick Controls.

Note: TableView will only load as many delegate items as needed to fill up the view. There is no guarantee that items outside the view will be loaded, although TableView will sometimes pre-load items for optimization reasons. Hence, a TableView with zero width or height might not load any delegate items at all.

Example Usage

C++ Models

The following example shows how to create a model from C++ with multiple columns:

#include <qqml.h>
#include <QAbstractTableModel>

class TableModel : public QAbstractTableModel
{
    Q_OBJECT
    QML_ELEMENT
    QML_ADDED_IN_VERSION(1, 1)

public:
    int rowCount(const QModelIndex & = QModelIndex()) const override
    {
        return 200;
    }

    int columnCount(const QModelIndex & = QModelIndex()) const override
    {
        return 200;
    }

    QVariant data(const QModelIndex &index, int role) const override
    {
        switch (role) {
            case Qt::DisplayRole:
                return QString("%1, %2").arg(index.column()).arg(index.row());
            default:
                break;
        }

        return QVariant();
    }

    QHash<int, QByteArray> roleNames() const override
    {
        return { {Qt::DisplayRole, "display"} };
    }
};

And then how to use it from QML:

import QtQuick
import TableModel 0.1

TableView {
    anchors.fill: parent
    columnSpacing: 1
    rowSpacing: 1
    clip: true

    model: TableModel {}

    delegate: Rectangle {
        implicitWidth: 100
        implicitHeight: 50
        Text {
            text: display
        }
    }
}

QML Models

For prototyping and displaying very simple data (from a web API, for example), TableModel can be used:

import QtQuick
import Qt.labs.qmlmodels

TableView {
    anchors.fill: parent
    columnSpacing: 1
    rowSpacing: 1
    clip: true

    model: TableModel {
        TableModelColumn { display: "name" }
        TableModelColumn { display: "color" }

        rows: [
            {
                "name": "cat",
                "color": "black"
            },
            {
                "name": "dog",
                "color": "brown"
            },
            {
                "name": "bird",
                "color": "white"
            }
        ]
    }

    delegate: Rectangle {
        implicitWidth: 100
        implicitHeight: 50
        border.width: 1

        Text {
            text: display
            anchors.centerIn: parent
        }
    }
}

Reusing items

TableView recycles delegate items by default, instead of instantiating from the delegate whenever new rows and columns are flicked into view. This approach gives a huge performance boost, depending on the complexity of the delegate.

When an item is flicked out, it moves to the reuse pool, which is an internal cache of unused items. When this happens, the TableView::pooled signal is emitted to inform the item about it. Likewise, when the item is moved back from the pool, the TableView::reused signal is emitted.

Any item properties that come from the model are updated when the item is reused. This includes index, row, and column, but also any model roles.

Note: Avoid storing any state inside a delegate. If you do, reset it manually on receiving the TableView::reused signal.

If an item has timers or animations, consider pausing them on receiving the TableView::pooled signal. That way you avoid using the CPU resources for items that are not visible. Likewise, if an item has resources that cannot be reused, they could be freed up.

If you don't want to reuse items or if the delegate cannot support it, you can set the reuseItems property to false.

Note: While an item is in the pool, it might still be alive and respond to connected signals and bindings.

The following example shows a delegate that animates a spinning rectangle. When it is pooled, the animation is temporarily paused:

Component {
    id: tableViewDelegate
    Rectangle {
        implicitWidth: 100
        implicitHeight: 50

        TableView.onPooled: rotationAnimation.pause()
        TableView.onReused: rotationAnimation.resume()

        Rectangle {
            id: rect
            anchors.centerIn: parent
            width: 40
            height: 5
            color: "green"

            RotationAnimation {
                id: rotationAnimation
                target: rect
                duration: (Math.random() * 2000) + 200
                from: 0
                to: 359
                running: true
                loops: Animation.Infinite
            }
        }
    }
}

Row heights and column widths

When a new column is flicked into view, TableView will determine its width by calling the columnWidthProvider. If set, this function will alone decide the width of the column. Otherwise, it will check if an explicit width has been set with setColumnWidth(). If not, implicitColumnWidth() will be used. The implicit width of a column is the same as the largest implicit width found among the currently loaded delegate items in that column. Trying to set an explicit width directly on a delegate has no effect, and will be ignored and overwritten. The same logic also applies to row heights.

An implementation of a columnWidthProvider that is equivalent to the default logic would be:

columnWidthProvider: function(column) {
    let w = explicitColumnWidth(column)
    if (w >= 0)
        return w;
    return implicitColumnWidth(column)
}

Once the column width is resolved, all other items in the same column are resized to this width, including any items that are flicked into the view at a later point.

Note: The resolved width of a column is discarded when the whole column is flicked out of the view, and is recalculated again if it's flicked back in. This means that if the width depends on the implicitColumnWidth(), the calculation can be different each time, depending on which row you're at when the column enters (since implicitColumnWidth() only considers the delegate items that are currently loaded). To avoid this, you should use a columnWidthProvider, or ensure that all the delegate items in the same column have the same implicitWidth.

If you change the values that a rowHeightProvider or a columnWidthProvider return for rows and columns inside the viewport, you must call forceLayout. This informs TableView that it needs to use the provider functions again to recalculate and update the layout.

Since Qt 5.13, if you want to hide a specific column, you can return 0 from the columnWidthProvider for that column. Likewise, you can return 0 from the rowHeightProvider to hide a row. If you return a negative number, TableView will fall back to calculate the size based on the delegate items.

Note: The size of a row or column should be a whole number to avoid sub-pixel alignment of items.

The following example shows how to set a simple columnWidthProvider together with a timer that modifies the values the function returns. When the array is modified, forceLayout is called to let the changes take effect:

TableView {
    id: tableView

    property var columnWidths: [100, 50, 80, 150]
    columnWidthProvider: function (column) { return columnWidths[column] }

    Timer {
        running: true
        interval: 2000
        onTriggered: {
            tableView.columnWidths[2] = 150
            tableView.forceLayout();
        }
    }
}

Editing cells

You can let the user edit table cells by providing an edit delegate. The edit delegate will be instantiated according to the editTriggers, which by default is when the user double taps on a cell, or presses e.g Qt::Key_Enter or Qt::Key_Return. The edit delegate is set using TableView::editDelegate, which is an attached property that you set on the delegate. The following snippet shows how to do that:

    TableView {
        id: tableView
        anchors.fill: parent
        clip: true

        model: TableModel {
            TableModelColumn { display: "name" }
            rows: [ { "name": "Harry" }, { "name": "Hedwig" } ]
        }

        selectionModel: ItemSelectionModel {}

        delegate: Rectangle {
            implicitWidth: 100
            implicitHeight: 50

            Text {
                anchors.centerIn: parent
                text: display
            }

            TableView.editDelegate: TextField {
                anchors.fill: parent
                text: display
                horizontalAlignment: TextInput.AlignHCenter
                verticalAlignment: TextInput.AlignVCenter
                Component.onCompleted: selectAll()

                TableView.onCommit: {
                    display = text
                    // 'display = text' is short-hand for:
                    // let index = TableView.view.index(row, column)
                    // TableView.view.model.setData(index, text, Qt.DisplayRole)
                }
            }
        }
    }

If the user presses Qt::Key_Enter or Qt::Key_Return while the edit delegate is active, TableView will emit the TableView::commit signal to the edit delegate, so that it can write back the changed data to the model.

Note: In order for a cell to be editable, the model needs to override QAbstractItemModel::flags(), and return Qt::ItemIsEditable. This flag is not enabled in QAbstractItemModel by default. The override could for example look like this:

Qt::ItemFlags QAbstractItemModelSubClass::flags(const QModelIndex &index) const override
{
    Q_UNUSED(index)
    return Qt::ItemIsSelectable | Qt::ItemIsEnabled | Qt::ItemIsEditable;
}

If the TableView delegate has a property required property bool editing defined, it will be set to true for the delegate being edited. See the documentation for editDelegate for an example on how to use it.

Overlays and underlays

All new items that are instantiated from the delegate are parented to the contentItem with the z value, 1. You can add your own items inside the Tableview, as child items of the Flickable. By controlling their z value, you can make them be on top of or underneath the table items.

Here is an example that shows how to add some text on top of the table, that moves together with the table as you flick:

TableView {
    id: tableView

    topMargin: header.implicitHeight

    Text {
        id: header
        text: "A table header"
    }
}

Here is another example that shows how to create an overlay item that stays on top of a particular cell. This requires a bit more code, since the location of a cell will change if the user, for example, is resizing a column in front of it.

Rectangle {
    id: overlay
    width: 20
    height: 20
    radius: 10
    color: "blue"

    z: 10
    parent: tableView.contentItem

    Connections {
        target: tableView
        function onLayoutChanged() {
            let item = tableView.itemAtCell(5, 5)
            let insideViewport = item !== null

            overlay.visible = insideViewport
            if (insideViewport) {
                overlay.x = item.x
                overlay.y = item.y
            }
        }
    }
}

You could also parent the overlay directly to the cell instead of the contentItem. But doing so will be fragile since the cell is unloaded or reused whenever it's flicked out of the viewport.

Selecting items

You can add selection support to TableView by assigning an ItemSelectionModel to the selectionModel property. It will then use this model to control which delegate items should be shown as selected, and which item should be shown as current. You can set selectionBehavior to control if the user should be allowed to select individual cells, rows, or columns.

To find out whether a delegate is selected or current, declare the following properties:

delegate: Item {
    required property bool selected
    required property bool current
    // ...
}

Note: the selected and current properties must be defined as required. This will inform TableView that it should take responsibility for updating their values. If not, they will simply be ignored. See also Required Properties.

The following snippet shows how an application can render the delegate differently depending on the selected property:

    TableView {
        id: tableView
        anchors.fill: parent
        clip: true

        model: TableModel {
            TableModelColumn { display: "name" }
            rows: [ { "name": "Harry" }, { "name": "Hedwig" } ]
        }

        selectionModel: ItemSelectionModel {}

        delegate: Rectangle {
            implicitWidth: 100
            implicitHeight: 30
            color: selected ? "blue" : "lightgray"

            required property bool selected

            Text { text: display }
        }
    }

The currentRow and currentColumn properties can also be useful if you need to render a delegate differently depending on if it lies on the same row or column as the current item.

Note: Qt Quick Controls offers a SelectionRectangle that can be used to let the user select cells.

Note: By default, a cell will become current, and any selections will be removed, when the user taps on it. If such default tap behavior is not wanted (e.g if you use custom pointer handlers inside your delegate), you can set pointerNavigationEnabled to false.

Keyboard navigation

In order to support keyboard navigation, you need to assign an ItemSelectionModel to the selectionModel property. TableView will then use this model to manipulate the model's currentIndex.

It's the responsibility of the delegate to render itself as current. You can do this by adding a property required property bool current to it, and let the appearance depend on its state. The current property's value is set by the TableView. You can also disable keyboard navigation fully (in case you want to implement your own key handlers) by setting keyNavigationEnabled to false.

The following example demonstrates how you can use keyboard navigation together with current and selected properties:

ApplicationWindow {
    width: 800
    height: 600
    visible: true
    ScrollView {
        anchors.fill: parent
        TableView {
            id: tableView
            clip: true
            interactive: true
            rowSpacing: 1
            columnSpacing: 1
            model: TableModel {
                TableModelColumn { display: "checked" }
                TableModelColumn { display: "amount" }
                TableModelColumn { display: "fruitType" }
                TableModelColumn { display: "fruitName" }
                TableModelColumn { display: "fruitPrice" }

                rows: [
                    {
                        checked: false,
                        amount: 1,
                        fruitType: "Apple",
                        fruitName: "Granny Smith",
                        fruitPrice: 1.50
                    },
                    {
                        checked: true,
                        amount: 4,
                        fruitType: "Orange",
                        fruitName: "Navel",
                        fruitPrice: 2.50
                    },
                    {
                        checked: false,
                        amount: 1,
                        fruitType: "Banana",
                        fruitName: "Cavendish",
                        fruitPrice: 3.50
                    }
                ]
            }
            selectionModel: ItemSelectionModel {}
            delegate: Rectangle {
                implicitWidth: 100
                implicitHeight: 50
                required property bool selected
                required property bool current
                border.width: current ? 2 : 0
                color: selected ? "lightblue" : palette.base
                Text{
                    text: model.display
                    padding: 12
                }
            }
        }
    }
    SelectionRectangle {
        target: tableView
    }
}

Copy and paste

Implementing copy and paste operations for a TableView usually also includes using a QUndoStack (or some other undo/redo framework). The QUndoStack can be used to store the different operations done on the model, like adding or removing rows, or pasting data from the clipboard, with a way to undo it again later. However, an accompanying QUndoStack that describes the possible operations, and how to undo them, should be designed according to the needs of the model and the application. As such, TableView doesn't offer a built-in API for handling copy and paste.

The following snippet can be used as a reference for how to add copy and paste support to your model and TableView. It uses the existing mime data API in QAbstractItemModel, together with QClipboard. The snippet will work as it is, but can also be extended to use a QUndoStack.

// Inside your C++ QAbstractTableModel subclass:

Q_INVOKABLE void copyToClipboard(const QModelIndexList &indexes) const
{
    QGuiApplication::clipboard()->setMimeData(mimeData(indexes));
}

Q_INVOKABLE bool pasteFromClipboard(const QModelIndex &targetIndex)
{
    const QMimeData *mimeData = QGuiApplication::clipboard()->mimeData();
    // Consider using a QUndoCommand for the following call. It should store
    // the (mime) data for the model items that are about to be overwritten, so
    // that a later call to undo can revert it.
    return dropMimeData(mimeData, Qt::CopyAction, -1, -1, targetIndex);
}

The two functions can, for example, be used from QML like this:

TableView {
    id: tableView
    model: tableModel
    selectionModel: ItemSelectionModel {}

    Shortcut {
       sequence: StandardKey.Copy
       onActivated: {
           let indexes = tableView.selectionModel.selectedIndexes
           tableView.model.copyToClipboard(indexes)
       }
    }

    Shortcut {
       sequence: StandardKey.Paste
       onActivated: {
           let targetIndex = tableView.selectionModel.currentIndex
           tableView.model.pasteFromClipboard(targetIndex)
       }
    }
}

See also TableView::editDelegate, TableView::commit, editTriggers, edit(), closeEditor(), layoutChanged(), QAbstractItemModel::mimeData(), QAbstractItemModel::dropMimeData(), QUndoStack, QUndoCommand, and QClipboard.

Property Documentation

alternatingRows : bool

This property controls whether the background color of the rows should alternate. The default value is style dependent.

Note: This property is only a hint, and might therefore not be respected by custom delegates. It's recommended that a delegate alternates between palette.base and palette.alternateBase when this hint is true, so that the colors can be set from outside of the delegate. For example:

background: Rectangle {
    color: control.row === control.tableView.currentRow
           ? control.palette.highlight
           : (control.tableView.alternatingRows && control.row % 2 !== 0
           ? control.palette.alternateBase
           : control.palette.base)
}

animate : bool [since 6.4]

This property can be set to control if TableView should animate the contentItem (contentX and contentY). It is used by positionViewAtCell(), and when navigating the current index with the keyboard. The default value is true.

If set to false, any ongoing animation will immediately stop.

Note: This property is only a hint. TableView might choose to position the content item without an animation if, for example, the target cell is not loaded. However, if set to false, animations will always be off.

This property was introduced in Qt 6.4.

See also positionViewAtCell().


bottomRow : int

This property holds the bottom-most row that is currently visible inside the view.

See also leftColumn, rightColumn, and topRow.


columnSpacing : real

This property holds the spacing between the columns.

The default value is 0.


columnWidthProvider : var

This property can hold a function that returns the column width for each column in the model. It is called whenever TableView needs to know the width of a specific column. The function takes one argument, column, for which the TableView needs to know the width.

Since Qt 5.13, if you want to hide a specific column, you can return 0 width for that column. If you return a negative number, TableView calculates the width based on the delegate items.

Note: The columnWidthProvider will usually be called two times when a column is about to load (or when doing layout). First, to know if the column is visible and should be loaded. And second, to determine the width of the column after all items have been loaded. If you need to calculate the column width based on the size of the delegate items, you need to wait for the second call, when all the items have been loaded. You can check for this by calling isColumnLoaded(column), and simply return -1 if that is not yet the case.

See also rowHeightProvider, isColumnLoaded(), and Row heights and column widths.


columns : int [read-only]

This property holds the number of columns in the table.

Note: columns is usually equal to the number of columns in the model, but can temporarily differ until all pending model changes have been processed.

If the model is a list, columns will be 1.

This property is read only.


contentHeight : real

This property holds the table height required to accommodate the number of rows in the data model. This is usually not the same as the height of the view, which means that the table's height could be larger or smaller than the viewport height. As a TableView cannot always know the exact height of the table without loading all rows in the model, the contentHeight is usually an estimate based on the initially loaded table.

If you know what the height of the table will be, assign a value to contentHeight, to avoid unnecessary calculations and updates to the TableView.

See also contentWidth and rowHeightProvider.


contentWidth : real

This property holds the table width required to accommodate the number of columns in the model. This is usually not the same as the width of the view, which means that the table's width could be larger or smaller than the viewport width. As a TableView cannot always know the exact width of the table without loading all columns in the model, the contentWidth is usually an estimate based on the initially loaded table.

If you know what the width of the table will be, assign a value to contentWidth, to avoid unnecessary calculations and updates to the TableView.

See also contentHeight and columnWidthProvider.


currentColumn : int [read-only]

This read-only property holds the column in the view that contains the item that is current. If no item is current, it will be -1.

Note: In order for TableView to report what the current column is, you need to assign an ItemSelectionModel to selectionModel.

See also currentRow, selectionModel, and Selecting items.


currentRow : int [read-only]

This read-only property holds the row in the view that contains the item that is current. If no item is current, it will be -1.

Note: In order for TableView to report what the current row is, you need to assign an ItemSelectionModel to selectionModel.

See also currentColumn, selectionModel, and Selecting items.


delegate : Component

The delegate provides a template defining each cell item instantiated by the view. The model index is exposed as an accessible index property. The same applies to row and column. Properties of the model are also available depending upon the type of Data Model.

A delegate should specify its size using implicitWidth and implicitHeight. The TableView lays out the items based on that information. Explicit width or height settings are ignored and overwritten.

Inside the delegate, you can optionally add one or more of the following properties. TableView modifies the values of these properties to inform the delegate which state it's in. This can be used by the delegate to render itself differently according on its own state.

  • required property bool current - true if the delegate is current.
  • required property bool selected - true if the delegate is selected.
  • required property bool editing - true if the delegate is being edited.

The following example shows how to use these properties:

delegate: Rectangle {
    required property bool current
    required property bool selected
    border.width: current ? 1 : 0
    color: selected ? palette.highlight : palette.base
}

Note: Delegates are instantiated as needed and may be destroyed at any time. They are also reused if the reuseItems property is set to true. You should therefore avoid storing state information in the delegates.

See also Row heights and column widths, Reusing items, and Required Properties.


editTriggers : enumeration [default: TableView.DoubleTapped | TableView.EditKeyPressed., since 6.5]

This property holds the different ways the user can start to edit a cell. It can be a combination of the following values:

ConstantDescription
TableView.NoEditTriggers- the user cannot trigger editing of cells. When this value is set, TableView will neither open or close the edit delegate as a response to any user interaction. But the application can call edit() and closeEditor() manually.
TableView.SingleTapped- the user can edit a cell by single tapping it.
TableView.DoubleTapped- the user can edit a cell by double tapping it.
TableView.SelectedTapped- the user can edit a selected cell by tapping it.
TableView.EditKeyPressed- the user can edit the current cell by pressing one of the edit keys. The edit keys are decided by the OS, but are normally Qt::Key_Enter and Qt::Key_Return.
TableView.AnyKeyPressed- the user can edit the current cell by pressing any key, other than the cell navigation keys. The pressed key is also sent to the focus object inside the edit delegate.

For TableView.SelectedTapped, TableView.EditKeyPressed, and TableView.AnyKeyPressed to have any effect, TableView needs to have a selection model assigned, since they depend on a current index being set. To be able to receive any key events at all, TableView will also need to have QQuickItem::activeFocus.

When editing a cell, the user can press Qt::Key_Tab or Qt::Key_Backtab to commit the data, and move editing to the next cell. This behavior can be disabled by setting QQuickItem::activeFocusOnTab on TableView to false.

Note: In order for a cell to be editable, the delegate needs an edit delegate attached, and the model needs to return Qt::ItemIsEditable from QAbstractItemModel::flags() (exemplified underneath). If you still cannot edit a cell after activating one of the specified triggers, you can, as a help, try to call edit() explicitly (e.g from a Button/TapHandler). Doing so will print out a warning explaining why the cell cannot be edited.

Qt::ItemFlags QAbstractItemModelSubClass::flags(const QModelIndex &index) const override
{
    Q_UNUSED(index)
    return Qt::ItemIsSelectable | Qt::ItemIsEnabled | Qt::ItemIsEditable;
}

This property was introduced in Qt 6.5.

See also TableView::editDelegate, TableView::commit, and Editing cells.


keyNavigationEnabled : bool [since 6.4]

This property can be set to control if the user should be able to change the current index using the keyboard. The default value is true.

Note: In order for TableView to support keyboard navigation, you need to assign an ItemSelectionModel to selectionModel.

This property was introduced in Qt 6.4.

See also Keyboard navigation, selectionModel, selectionBehavior, pointerNavigationEnabled, and interactive.


leftColumn : int

This property holds the leftmost column that is currently visible inside the view.

See also rightColumn, topRow, and bottomRow.


model : model

This property holds the model that provides data for the table.

The model provides the set of data that is used to create the items in the view. Models can be created directly in QML using TableModel, ListModel, ObjectModel, or provided by a custom C++ model class. The C++ model must be a subclass of QAbstractItemModel or a simple list.

See also Data Models.


pointerNavigationEnabled : bool [since 6.4]

This property can be set to control if the user should be able to change the current index using mouse or touch. The default value is true.

This property was introduced in Qt 6.4.

See also selectionModel, keyNavigationEnabled, and interactive.


resizableColumns : bool [since 6.5]

This property holds whether the user is allowed to resize columns by dragging between the cells. The default value is false.

This property was introduced in Qt 6.5.


resizableRows : bool [since 6.5]

This property holds whether the user is allowed to resize rows by dragging between the cells. The default value is false.

This property was introduced in Qt 6.5.


reuseItems : bool

This property holds whether or not items instantiated from the delegate should be reused. If set to false, any currently pooled items are destroyed.

See also Reusing items, TableView::pooled, and TableView::reused.


rightColumn : int

This property holds the rightmost column that is currently visible inside the view.

See also leftColumn, topRow, and bottomRow.


rowHeightProvider : var

This property can hold a function that returns the row height for each row in the model. It is called whenever TableView needs to know the height of a specific row. The function takes one argument, row, for which the TableView needs to know the height.

Since Qt 5.13, if you want to hide a specific row, you can return 0 height for that row. If you return a negative number, TableView calculates the height based on the delegate items.

Note: The rowHeightProvider will usually be called two times when a row is about to load (or when doing layout). First, to know if the row is visible and should be loaded. And second, to determine the height of the row after all items have been loaded. If you need to calculate the row height based on the size of the delegate items, you need to wait for the second call, when all the items have been loaded. You can check for this by calling isRowLoaded(row), and simply return -1 if that is not yet the case.

See also rowHeightProvider, isRowLoaded(), and Row heights and column widths.


rowSpacing : real

This property holds the spacing between the rows.

The default value is 0.


rows : int [read-only]

This property holds the number of rows in the table.

Note: rows is usually equal to the number of rows in the model, but can temporarily differ until all pending model changes have been processed.

This property is read only.


selectionBehavior : enumeration [since 6.4]

This property holds whether the user can select cells, rows or columns.

ConstantDescription
TableView.SelectionDisabledThe user cannot perform selections
TableView.SelectCells(Default value) The user can select individual cells
TableView.SelectRowsThe user can only select rows
TableView.SelectColumnsThe user can only select columns

This property was introduced in Qt 6.4.

See also Selecting items, selectionMode, selectionModel, and keyNavigationEnabled.


selectionMode : enumeration [since 6.6]

If selectionBehavior is set to TableView.SelectCells, this property holds whether the user can select one cell at a time, or multiple cells. If selectionBehavior is set to TableView.SelectRows, this property holds whether the user can select one row at a time, or multiple rows. If selectionBehavior is set to TableView.SelectColumns, this property holds whether the user can select one column at a time, or multiple columns.

The following modes are available:

ConstantDescription
TableView.SingleSelectionThe user can select a single cell, row or column.
TableView.ContiguousSelectionThe user can select a single contiguous block of cells. An existing selection can be made bigger or smaller by holding down the Shift modifier while selecting.
TableView.ExtendedSelection(Default value) The user can select multiple individual blocks of cells. An existing selection can be made bigger or smaller by holding down the Shift modifier while selecting. A new selection block can be started without clearing the current selection by holding down the Control modifier while selecting.

This property was introduced in Qt 6.6.

See also Selecting items, selectionBehavior, selectionModel, and keyNavigationEnabled.


selectionModel : ItemSelectionModel [since 6.2]

This property can be set to control which delegate items should be shown as selected, and which item should be shown as current. If the delegate has a required property bool selected defined, TableView will keep it in sync with the selection state of the corresponding model item in the selection model. If the delegate has a required property bool current defined, TableView will keep it in sync with selectionModel.currentIndex.

This property was introduced in Qt 6.2.

See also Selecting items, SelectionRectangle, keyNavigationEnabled, and pointerNavigationEnabled.


syncDirection : Qt::Orientations

If the syncView is set on a TableView, this property controls synchronization of flicking direction(s) for both tables. The default is Qt.Horizontal | Qt.Vertical, which means that if you flick either table in either direction, the other table is flicked the same amount in the same direction.

This property and syncView can be used to make two tableViews synchronize with each other smoothly in flicking regardless of the different overshoot/undershoot, velocity, acceleration/deceleration or rebound animation, and so on.

A typical use case is to make several headers flick along with the table.

See also syncView.


syncView : TableView

If this property of a TableView is set to another TableView, both the tables will synchronize with regard to flicking, column widths/row heights, and spacing according to syncDirection.

If syncDirection contains Qt.Horizontal, current tableView's column widths, column spacing, and horizontal flicking movement synchronizes with syncView's.

If syncDirection contains Qt.Vertical, current tableView's row heights, row spacing, and vertical flicking movement synchronizes with syncView's.

See also syncDirection.


topRow : int

This property holds the topmost row that is currently visible inside the view.

See also leftColumn, rightColumn, and bottomRow.


Attached Property Documentation

TableView.editDelegate : Component

This attached property holds the edit delegate. It's instantiated when editing begins, and parented to the delegate it edits. It supports the same required properties as the TableView delegate, including index, row and column. Properties of the model, like display and edit, are also available (depending on the role names exposed by the model).

Editing starts when the actions specified by editTriggers are met, and the current cell is editable.

Note: In order for a cell to be editable, the model needs to override QAbstractItemModel::flags(), and return Qt::ItemIsEditable.

You can also open and close the edit delegate manually by calling edit() and closeEditor(), respectively. The Qt::ItemIsEditable flag will then be ignored.

Editing ends when the user presses Qt::Key_Enter or Qt::Key_Return (and also Qt::Key_Tab or Qt::Key_Backtab, if TableView has QQuickItem::activeFocusOnTab set). In that case, the TableView::commit signal will be emitted, so that the edit delegate can respond by writing any modified data back to the model. If editing ends because of other reasons (e.g if the user presses Qt::Key_Escape), the signal will not be emitted. In any case will destruction() be emitted in the end.

While the edit delegate is showing, the cell underneath will still be visible, and therefore shine through if the edit delegate is translucent, or otherwise doesn't cover the whole cell. If this is not wanted, you can either let the root item of the edit delegate be a solid Rectangle, or hide some of the items inside the TableView delegate.. The latter can be done by defining a property required property bool editing inside it, that you bind to the visible property of some of the child items. The following snippet shows how to do that:

        delegate: Rectangle {
            implicitWidth: 100
            implicitHeight: 50

            required property bool editing

            Text {
                id: textField
                anchors.fill: parent
                anchors.margins: 5
                text: display
                visible: !editing
            }

            TableView.editDelegate: TextField {
                x: textField.x
                y: textField.y
                width: textField.width
                height: textField.height
                text: display
                TableView.onCommit: display = text
            }
        }

When the edit delegate is instantiated, TableView will call QQuickItem::forceActiveFocus() on it. If you want active focus to be set on a child of the edit delegate instead, let the edit delegate be a FocusScope.

See also editTriggers, TableView::commit, edit(), closeEditor(), and Editing cells.


TableView.view : TableView

This attached property holds the view that manages the delegate instance. It is attached to each instance of the delegate.


Signal Documentation

[since 6.5] layoutChanged()

This signal is emitted whenever the layout of the loaded rows and columns has potentially changed. This will especially be the case when forceLayout() is called, but also when e.g resizing a row or a column, or when a row or column have entered or left the viewport.

This signal can be used to for example update the geometry of overlays.

Note: The corresponding handler is onLayoutChanged.

This signal was introduced in Qt 6.5.

See also forceLayout() and Overlays and underlays.


Attached Signal Documentation

commit()

This signal is emitted by the edit delegate

This attached signal is emitted when the edit delegate is active, and the user presses Qt::Key_Enter or Qt::Key_Return. It will also be emitted if TableView has QQuickItem::activeFocusOnTab set, and the user presses Qt::Key_Tab or Qt::Key_Backtab.

This signal will not be emitted if editing ends because of reasons other than the ones mentioned. This includes e.g if the user presses Qt::Key_Escape, taps outside the delegate, the row or column being edited is deleted, or if the application calls closeEditor().

Upon receiving the signal, the edit delegate should write any modified data back to the model.

Note: This property should be attached to the edit delegate, and not to the delegate.

Note: The corresponding handler is onCommit.

See also TableView::editDelegate, editTriggers, and Editing cells.


pooled()

This signal is emitted after an item has been added to the reuse pool. You can use it to pause ongoing timers or animations inside the item, or free up resources that cannot be reused.

This signal is emitted only if the reuseItems property is true.

Note: The corresponding handler is onPooled.

See also Reusing items, reuseItems, and reused.


reused()

This signal is emitted after an item has been reused. At this point, the item has been taken out of the pool and placed inside the content view, and the model properties such as index, row, and column have been updated.

Other properties that are not provided by the model does not change when an item is reused. You should avoid storing any state inside a delegate, but if you do, manually reset that state on receiving this signal.

This signal is emitted when the item is reused, and not the first time the item is created.

This signal is emitted only if the reuseItems property is true.

Note: The corresponding handler is onReused.

See also Reusing items, reuseItems, and pooled.


Method Documentation

[since 6.4] point cellAtIndex(QModelIndex modelIndex)

Returns the cell in the view that maps to modelIndex in the model. Convenience function for doing:

Qt.point(columnAtIndex(modelIndex), rowAtIndex(modelIndex))

A cell is simply a point that combines row and column into a single type.

Note: that point.x will map to the column, and point.y will map to the row.

This method was introduced in Qt 6.4.


Point cellAtPosition(point position, bool includeSpacing)

Returns the cell at the given position in the table. position should be relative to the contentItem. If no loaded cell intersects with position, the return value will be point(-1, -1).

If includeSpacing is set to true, a cell's bounding box will be considered to include half the adjacent rowSpacing and columnSpacing on each side. The default value is false.

Note: A Input Handler attached to a TableView installs itself on the contentItem rather than the view. So the position reported by the handler can be used directly in a call to this function without any mapping.

See also columnSpacing and rowSpacing.


Point cellAtPosition(real x, real y, bool includeSpacing)

Convenience for calling cellAtPosition(Qt.point(x, y), includeSpacing).


clearColumnWidths()

Clears all the column widths set with setColumnWidth().

Note: If a syncView is set, together with a Qt.Horizontal syncDirection, the sync view will control the column widths. Therefore, in that case, any call to this function will be forwarded to the sync view instead.

See also setColumnWidth(), clearRowHeights(), and Row heights and column widths.


clearRowHeights()

Clears all the row heights set with setRowHeight().

Note: If a syncView is set, together with a Qt.Vertical syncDirection, the sync view will control the row heights. Therefore, in that case, any call to this function will be forwarded to the sync view instead.

See also setRowHeight(), clearColumnWidths(), and Row heights and column widths.


[since 6.5] closeEditor()

If the user is editing a cell, calling this function will stop the editing, and destroy the edit delegate instance.

This method was introduced in Qt 6.5.

See also edit(), TableView::editDelegate, and Editing cells.


[since 6.4] int columnAtIndex(QModelIndex modelIndex)

Returns the column in the view that maps to modelIndex in the model.

This method was introduced in Qt 6.4.

See also rowAtIndex() and index().


[since 6.2] real columnWidth(int column)

Returns the width of the given column. If the column is not loaded (and therefore not visible), the return value will be -1.

This method was introduced in Qt 6.2.

See also setColumnWidth(), columnWidthProvider, implicitColumnWidth(), isColumnLoaded(), and Row heights and column widths.


[since 6.5] edit(QModelIndex modelIndex)

This function starts an editing session for the cell that represents modelIndex. If the user is already editing another cell, that session ends.

Normally you can specify the different ways of starting an edit session by using editTriggers instead. If that isn't sufficient, you can use this function. To take full control over cell editing and keep TableView from interfering, set editTriggers to TableView.NoEditTriggers.

Note: The current index in the selection model will also change to modelIndex.

This method was introduced in Qt 6.5.

See also closeEditor(), editTriggers, TableView::editDelegate, and Editing cells.


qreal explicitColumnWidth(int column)

Returns the width of the column set with setColumnWidth(). This width might differ from the actual width of the column, if a columnWidthProvider is in use. To get the actual width of a column, use columnWidth().

A return value equal to 0 means that the column has been told to hide. A return value equal to -1 means that no explicit width has been set for the column.

Note: If a syncView is set, together with a Qt.Horizontal syncDirection, the sync view will control the column widths. Therefore, in that case, any call to this function will be forwarded to the sync view instead.

See also setColumnWidth(), columnWidth(), and Row heights and column widths.


qreal explicitRowHeight(int row)

Returns the height of the row set with setRowHeight(). This height might differ from the actual height of the column, if a rowHeightProvider is in use. To get the actual height of a row, use rowHeight().

A return value equal to 0 means that the row has been told to hide. A return value equal to -1 means that no explicit height has been set for the row.

Note: If a syncView is set, together with a Qt.Vertical syncDirection, the sync view will control the row heights. Therefore, in that case, any call to this function will be forwarded to the sync view instead.

See also setRowHeight(), rowHeight(), and Row heights and column widths.


forceLayout()

Responding to changes in the model are batched so that they are handled only once per frame. This means the TableView delays showing any changes while a script is being run. The same is also true when changing properties, such as rowSpacing or leftMargin.

This method forces the TableView to immediately update the layout so that any recent changes take effect.

Calling this function re-evaluates the size and position of each visible row and column. This is needed if the functions assigned to rowHeightProvider or columnWidthProvider return different values than what is already assigned.


[since 6.2] real implicitColumnWidth(int column)

Returns the implicit width of the given column. If the column is not loaded (and therefore not visible), the return value will be -1.

The implicit width of a column is the largest implicitWidth found among the currently loaded delegate items inside that column. Widths returned by the columnWidthProvider will not be taken into account.

This method was introduced in Qt 6.2.

See also columnWidthProvider, columnWidth(), isColumnLoaded(), and Row heights and column widths.


[since 6.2] real implicitRowHeight(int row)

Returns the implicit height of the given row. If the row is not loaded (and therefore not visible), the return value will be -1.

The implicit height of a row is the largest implicitHeight found among the currently loaded delegate items inside that row. Heights returned by the rowHeightProvider will not be taken into account.

This method was introduced in Qt 6.2.

See also rowHeightProvider, rowHeight(), isRowLoaded(), and Row heights and column widths.


[since 6.4.3] QModelIndex index(int row, int column)

Returns the QModelIndex that maps to row and column in the view.

row and column should be the row and column in the view (table row and table column), and not a row and column in the model. For a plain TableView, this is equivalent of calling model.index(row, column). But for a subclass of TableView, like TreeView, where the data model is wrapped inside an internal proxy model that flattens the tree structure into a table, you need to use this function to resolve the model index.

This method was introduced in Qt 6.4.3.

See also rowAtIndex() and columnAtIndex().


[since 6.2] bool isColumnLoaded(int column)

Returns true if the given column is loaded.

A column is loaded when TableView has loaded the delegate items needed to show the column inside the view. This also usually means that the column is visible for the user, but not always.

This function can be used whenever you need to iterate over the delegate items for a column, e.g from a columnWidthProvider, to be sure that the delegate items are available for iteration.

This method was introduced in Qt 6.2.


[since 6.2] bool isRowLoaded(int row)

Returns true if the given row is loaded.

A row is loaded when TableView has loaded the delegate items needed to show the row inside the view. This also usually means that the row is visible for the user, but not always.

This function can be used whenever you need to iterate over the delegate items for a row, e.g from a rowHeightProvider, to be sure that the delegate items are available for iteration.

This method was introduced in Qt 6.2.


Item itemAtCell(point cell)

Returns the delegate item at cell if loaded, otherwise null.

Note: only the items that are visible in the view are normally loaded. As soon as a cell is flicked out of the view, the item inside will either be unloaded or placed in the recycle pool. As such, the return value should never be stored.


[since 6.5] Item itemAtIndex(QModelIndex index)

Returns the instantiated delegate item for the cell that represents index. If the item is not loaded, the value will be null.

Note: only the items that are visible in the view are normally loaded. As soon as a cell is flicked out of the view, the item inside will either be unloaded or placed in the recycle pool. As such, the return value should never be stored.

Note: If the model is not a QAbstractItemModel, you can also use itemAtCell(Qt.point(column, row)). But be aware that point.x maps to columns and point.y maps to rows.

This method was introduced in Qt 6.5.


[since 6.4] QModelIndex modelIndex(point cell)

Convenience function for doing:

modelIndex(cell.y, cell.x)

A cell is simply a point that combines row and column into a single type.

Note: point.x will map to the column, and point.y will map to the row.

This method was introduced in Qt 6.4.


positionViewAtCell(point cell, PositionMode mode, point offset, rect subRect)

Positions contentX and contentY such that cell is at the position specified by mode. mode can be an or-ed combination of the following:

ConstantDescription
TableView.AlignLeftPosition the cell at the left of the view.
TableView.AlignHCenterPosition the cell at the horizontal center of the view.
TableView.AlignRightPosition the cell at the right of the view.
TableView.AlignTopPosition the cell at the top of the view.
TableView.AlignVCenterPosition the cell at the vertical center of the view.
TableView.AlignBottomPosition the cell at the bottom of the view.
TableView.AlignCenterThe same as (TableView.AlignHCenter | TableView.AlignVCenter)
TableView.VisibleIf any part of the cell is visible then take no action. Otherwise move the content item so that the entire cell becomes visible.
TableView.ContainIf the entire cell is visible then take no action. Otherwise move the content item so that the entire cell becomes visible. If the cell is bigger than the view, the top-left part of the cell will be preferred.

If no vertical alignment is specified, vertical positioning will be ignored. The same is true for horizontal alignment.

Optionally, you can specify offset to move contentX and contentY an extra number of pixels beyond the target alignment. E.g if you want to position the view so that cell [10, 10] ends up at the top-left corner with a 5px margin, you could do:

positionViewAtCell(Qt.point(10, 10), TableView.AlignLeft | TableView.AlignTop, Qt.point(-5, -5))

As of Qt 6.4, you can specify a subRect to position on a rectangle inside the cell, rather than on the bounding rectangle of the whole cell. This can be useful if the cell is e.g larger than the view, and you want to ensure that a specific part of it is visible. The subRect needs to be valid to be taken into consideration.

Note: It is not recommended to use contentX or contentY to position the view at a particular cell. This is unreliable since removing items from the start of the table does not cause all other items to be repositioned. TableView can also sometimes place rows and columns at approximate positions to optimize for speed. The only exception is if the cell is already visible in the view, which can be checked upfront by calling itemAtCell().

Methods should only be called after the Component has completed. To position the view at startup, this method should be called by Component.onCompleted. For example, to position the view at the end:

Component.onCompleted: positionViewAtCell(Qt.point(columns - 1, rows - 1), TableView.AlignRight | TableView.AlignBottom)

Note: The second argument to this function used to be Qt.Alignment. For backwards compatibility, that enum can still be used. The change to use PositionMode was done in Qt 6.4.

See also animate.


positionViewAtColumn(int column, PositionMode mode, real offset, rect subRect)

Positions {Flickable::}{contentX} such that column is at the position specified by mode, offset and subRect.

Convenience method for calling

positionViewAtCell(Qt.point(column, 0), mode & Qt.AlignHorizontal_Mask, offset, subRect)

[since 6.5] positionViewAtIndex(QModelIndex index, PositionMode mode, point offset, rect subRect)

Positions the view such that index is at the position specified by mode, offset and subRect.

Convenience method for calling

positionViewAtRow(rowAtIndex(index), mode & Qt.AlignVertical_Mask, offset.y, subRect)
positionViewAtColumn(columnAtIndex(index), mode & Qt.AlignVertical_Mask, offset.x, subRect)

This method was introduced in Qt 6.5.


positionViewAtRow(int row, PositionMode mode, real offset, rect subRect)

Positions {Flickable::}{contentY} such that row is at the position specified by mode, offset and subRect.

Convenience method for calling

positionViewAtCell(Qt.point(0, row), mode & Qt.AlignVertical_Mask, offset, subRect)

[since 6.4] int rowAtIndex(QModelIndex modelIndex)

Returns the row in the view that maps to modelIndex in the model.

This method was introduced in Qt 6.4.

See also columnAtIndex() and index().


[since 6.2] real rowHeight(int row)

Returns the height of the given row. If the row is not loaded (and therefore not visible), the return value will be -1.

This method was introduced in Qt 6.2.

See also setRowHeight(), rowHeightProvider, implicitRowHeight(), isRowLoaded(), and Row heights and column widths.


setColumnWidth(int column, real size)

Sets the explicit column width of column column to size.

If you want to read back the values you set with this function, you should use explicitColumnWidth(). columnWidth() will return the actual size of the column, which can be different if a columnWidthProvider is set.

When TableView needs to resolve the width of column, it will first try to call the columnWidthProvider. Only if a provider is not set, will the widths set with this function be used by default. You can, however, call explicitColumnWidth() from within the provider, and if needed, moderate the values to e.g always be within a certain interval. The following snippet shows an example on how to do that:

columnWidthProvider: function(column) {
    let w = explicitColumnWidth(column)
    if (w >= 0)
        return Math.max(100, w);
    return implicitColumnWidth(column)
}

If size is equal to 0, the column will be hidden. If size is equal to -1, the column will be reset back to use implicitColumnWidth(). You are allowed to specify column sizes for columns that are outside the size of the model.

Note: The sizes you set will not be cleared if you change the model. To clear the sizes, you need to call clearColumnWidths() explicitly.

Note: If a syncView is set, together with a Qt.Horizontal syncDirection, the sync view will control the column widths. Therefore, in that case, any call to this function will be forwarded to the sync view instead.

Note: For models with lots of columns, using setColumnWidth() to set the widths for all the columns at start-up, can be suboptimal. This will consume start-up time and memory (for storing all the widths). A more scalable approach is to use a columnWidthProvider instead, or rely on the implicit width of the delegate. A columnWidthProvider will only be called on an as-needed basis, and will not be affected by the size of the model.

See also columnWidth(), explicitColumnWidth(), setRowHeight(), clearColumnWidths(), and Row heights and column widths.


setRowHeight(int row, real size)

Sets the explicit row height of row row to size.

If you want to read back the values you set with this function, you should use explicitRowHeight(). rowHeight() will return the actual height of the row, which can be different if a rowHeightProvider is set.

When TableView needs to resolve the height of row, it will first try to call the rowHeightProvider. Only if a provider is not set, will the heights set with this function be used by default. You can, however, call explicitRowHeight() from within the provider, and if needed, moderate the values to e.g always be within a certain interval. The following snippet shows an example on how to do that:

rowHeightProvider: function(row) {
    let h = explicitRowHeight(row)
    if (h >= 0)
        return Math.max(100, h);
    return implicitRowHeight(row)
}

If size is equal to 0, the row will be hidden. If size is equal to -1, the row will be reset back to use implicitRowHeight(). You are allowed to specify row sizes for rows that are outside the size of the model.

Note: The sizes you set will not be cleared if you change the model. To clear the sizes, you need to call clearRowHeights() explicitly.

Note: If a syncView is set, together with a Qt.Vertical syncDirection, the sync view will control the row heights. Therefore, in that case, any call to this function will be forwarded to the sync view instead.

Note: For models with lots of rows, using setRowHeight() to set the heights for all the rows at start-up, can be suboptimal. This will consume start-up time and memory (for storing all the heights). A more scalable approach is to use a rowHeightProvider instead, or rely on the implicit height of the delegate. A rowHeightProvider will only be called on an as-needed basis, and will not be affected by the size of the model.

See also rowHeight(), explicitRowHeight(), setColumnWidth(), and Row heights and column widths.


© 2024 The Qt Company Ltd. Documentation contributions included herein are the copyrights of their respective owners. The documentation provided herein is licensed under the terms of the GNU Free Documentation License version 1.3 as published by the Free Software Foundation. Qt and respective logos are trademarks of The Qt Company Ltd. in Finland and/or other countries worldwide. All other trademarks are property of their respective owners.