The Table
component is intended for presenting tabular
data organized in rows and columns. The Table
is one of
the most versatile components in Vaadin. Table cells can include text
or arbitrary UI components. You can easily implement editing of the table data,
for example clicking on a cell could change it to a text field for editing.
The data contained in a Table
is managed using the Data
Model of Vaadin (see Chapter 9, Binding Components to Data), through
the Container
interface of the
Table
. This makes it possible to bind a table directly
to a data source, such as a database query. Only the visible part of the table
is loaded into the browser and moving the visible window with the scrollbar loads
content from the server. While the data is being loaded, a tooltip will be
displayed that shows the current range and total number of items in the
table. The rows of the table are items in the container
and the columns are properties. Each table row (item) is
identified with an item identifier (IID), and each column
(property) with a property identifier (PID).
When creating a table, you first need to define columns with
addContainerProperty()
. This method comes in two
flavors. The simpler one takes the property ID of the column and uses it also
as the caption of the column. The more complex one allows differing PID and
header for the column. This may make, for example, internationalization of
table headers easier, because if a PID is internationalized, the
internationalization has to be used everywhere where the PID is used. The
complex form of the method also allows defining an icon for the column from a
resource. The "default value" parameter is used when new properties (columns)
are added to the table, to fill in the missing values. (This default has no
meaning in the usual case, such as below, where we add items after defining
the properties.)
/* Create the table with a caption. */ Table table = new Table("This is my Table"); /* Define the names and data types of columns. * The "default value" parameter is meaningless here. */ table.addContainerProperty("First Name", String.class, null); table.addContainerProperty("Last Name", String.class, null); table.addContainerProperty("Year", Integer.class, null); /* Add a few items in the table. */ table.addItem(new Object[] { "Nicolaus","Copernicus",new Integer(1473)}, new Integer(1)); table.addItem(new Object[] { "Tycho", "Brahe", new Integer(1546)}, new Integer(2)); table.addItem(new Object[] { "Giordano","Bruno", new Integer(1548)}, new Integer(3)); table.addItem(new Object[] { "Galileo", "Galilei", new Integer(1564)}, new Integer(4)); table.addItem(new Object[] { "Johannes","Kepler", new Integer(1571)}, new Integer(5)); table.addItem(new Object[] { "Isaac", "Newton", new Integer(1643)}, new Integer(6));
In this example, we used an increasing Integer
object
as the Item Identifier, given as the second parameter to
addItem()
. The actual rows are given simply as object
arrays, in the same order in which the properties were added. The objects must
be of the correct class, as defined in the
addContainerProperty()
calls.
Scalability of the Table
is largely dictated by
the container. The default IndexedContainer
is
relatively heavy and can cause scalability problems, for example, when
updating the values. Use of an optimized application-specific
container is recommended. Table does not have a limit for the number
of items and is just as fast with hundreds of thousands of items as
with just a few. With the current implementation of scrolling, there
is a limit of around 500 000 rows, depending on the browser and the pixel
height of rows.
The Table
allows selecting one or more items by
clicking them with the mouse. When the user selects an item, the IID of the
item will be set as the property of the table and a
ValueChangeEvent
is triggered. To enable selection,
you need to set the table selectable. You will also
need to set it as immediate in most cases, as we do
below, because without it, the change in the property will not be
communicated immediately to the server.
The following example shows how to enable the selection of items in a
Table
and how to handle
ValueChangeEvent
events that are caused by changes
in selection. You need to handle the event with the
valueChange()
method of the
Property.ValueChangeListener
interface.
// Allow selecting items from the table. table.setSelectable(true); // Send changes in selection immediately to server. table.setImmediate(true); // Shows feedback from selection. final Label current = new Label("Selected: -"); // Handle selection change. table.addListener(new Property.ValueChangeListener() { public void valueChange(ValueChangeEvent event) { current.setValue("Selected: " + table.getValue()); } });
If the user clicks on an already selected item, the selection will
deselected and the table property will have null
value. You can disable this behaviour by setting
setNullSelectionAllowed(false)
for the table.
The selection is the value of the table's property, so you can get it with
getValue()
. You can get it also from a reference
to the table itself. In single selection mode, the value is the item
identifier of the selected item or null
if no item
is selected. In multiple selection mode (see below), the value is a
Set
of item identifiers. Notice that the set is
unmodifiable, so you can not simply change it to change the selection.
A table can also be in multiselect mode, where a user can select multiple items by clicking them with left mouse button while holding the Ctrl key (or Meta key) pressed. If Ctrl is not held, clicking an item will select it and other selected items are deselected. The user can select a range by selecting an item, holding the Shift key pressed, and clicking another item, in which case all the items between the two are also selected. Multiple ranges can be selected by first selecting a range, then selecting an item while holding Ctrl, and then selecting another item with both Ctrl and Shift pressed.
The multiselect mode is enabled with the
setMultiSelect()
method of the
Select
interface of
Table
. Setting table in multiselect mode does
not implicitly set it as selectable, so it must
be set separately.
The setMultiSelectMode()
property affects the
control of multiple selection:
MultiSelectMode.DEFAULT
is the default
behaviour, which requires holding the Ctrl (or
Meta) key pressed while selecting items, while in
MultiSelectMode.SIMPLE
holding the
Ctrl key is not needed. In the simple mode, items can
only be deselected by clicking them.
The default style for Table
provides a table
with a scrollbar. The scrollbar is located at the right side of the
table and becomes visible when the number of items in the table
exceeds the page length, that is, the number of visible items. You can
set the page length with setPageLength()
.
Setting the page length to zero makes all the rows in a table visible, no matter how many rows there are. Notice that this also effectively disables buffering, as all the entire table is loaded to the browser at once. Using such tables to generate reports does not scale up very well, as there is some inevitable overhead in rendering a table with Ajax. For very large reports, generating HTML directly is a more scalable solution.
You can set the width of a column programmatically from the
server-side with setColumnWidth()
. The column
is identified by the property ID and the width is given in pixels.
The user can resize table columns by dragging the resize handle
between two columns. Resizing a table column causes a
ColumnResizeEvent
, which you can handle with a
Table.ColumnResizeListener
. The table must be
set in immediate mode if you want to receive the resize events
immediately, which is typical.
table.addListener(new Table.ColumnResizeListener() { public void columnResize(ColumnResizeEvent event) { // Get the new width of the resized column int width = event.getCurrentWidth(); // Get the property ID of the resized column String column = (String) event.getPropertyId(); // Do something with the information table.setColumnFooter(column, String.valueOf(width) + "px"); } }); // Must be immediate to send the resize events immediately table.setImmediate(true);
See Figure 5.48, “Resizing Columns” for a result after the columns of a table has been resized.
If setColumnReorderingAllowed(true)
is set,
the user can reorder table columns by dragging them with the mouse
from the column header,
When setColumnCollapsingAllowed(true)
is set,
the right side of the table header shows a drop-down list that allows
selecting which columns are shown. Collapsing columns is different
than hiding columns with setVisibleColumns()
,
which hides the columns completely so that they can not be made
visible (uncollapsed) from the user interface.
You can collapse columns programmatically with
setColumnCollapsed()
. Collapsing must be
enabled before collapsing columns with the method or it will throw an
IllegalAccessException
.
// Allow the user to collapse and uncollapse columns table.setColumnCollapsingAllowed(true); // Collapse this column programmatically try { table.setColumnCollapsed("born", true); } catch (IllegalAccessException e) { // Can't occur - collapsing was allowed above System.err.println("Something horrible occurred"); } // Give enough width for the table to accommodate the // initially collapsed column later table.setWidth("250px");
See Figure 5.49, “Collapsing Columns”.
If the table has undefined width, it minimizes its width to fit the width of the visible columns. If some columns are initially collapsed, the width of the table may not be enough to accomodate them later, which will result in an ugly horizontal scrollbar. You should consider giving the table enough width to accomodate columns uncollapsed by the user.
The cells of a Table
can contain any user
interface components, not just strings. If the rows are higher than
the row height defined in the default theme, you have to define the
proper row height in a custom theme.
When handling events for components inside a
Table
, such as for the
Button
in the example below, you usually need
to know the item the component belongs to. Components do not
themselves know about the table or the specific item in which a
component is contained. Therefore, the handling method must use some
other means for finding out the Item ID of the item. There are a few
possibilities. Usually the easiest way is to use the
setData()
method to attach an arbitrary
object to a component. You can subclass the component and include the
identity information there. You can also simply search the entire
table for the item with the component, although that solution may not
be so scalable.
The example below includes table rows with a
Label
in XHTML formatting mode, a multiline
TextField
, a CheckBox
,
and a Button
that shows as a link.
// Create a table and add a style to allow setting the row height in theme. final Table table = new Table(); table.addStyleName("components-inside"); /* Define the names and data types of columns. * The "default value" parameter is meaningless here. */ table.addContainerProperty("Sum", Label.class, null); table.addContainerProperty("Is Transferred", CheckBox.class, null); table.addContainerProperty("Comments", TextField.class, null); table.addContainerProperty("Details", Button.class, null); /* Add a few items in the table. */ for (int i=0; i<100; i++) { // Create the fields for the current table row Label sumField = new Label(String.format( "Sum is <b>$%04.2f</b><br/><i>(VAT incl.)</i>", new Object[] {new Double(Math.random()*1000)}), Label.CONTENT_XHTML); CheckBox transferredField = new CheckBox("is transferred"); // Multiline text field. This required modifying the // height of the table row. TextField commentsField = new TextField(); commentsField.setRows(3); // The Table item identifier for the row. Integer itemId = new Integer(i); // Create a button and handle its click. A Button does not // know the item it is contained in, so we have to store the // item ID as user-defined data. Button detailsField = new Button("show details"); detailsField.setData(itemId); detailsField.addListener(new Button.ClickListener() { public void buttonClick(ClickEvent event) { // Get the item identifier from the user-defined data. Integer itemId = (Integer)event.getButton().getData(); getWindow().showNotification("Link "+ itemId.intValue()+" clicked."); } }); detailsField.addStyleName("link"); // Create the table row. table.addItem(new Object[] {sumField, transferredField, commentsField, detailsField}, itemId); } // Show just three rows because they are so high. table.setPageLength(3);
The row height has to be set higher than the default with a style rule such as the following:
/* Table rows contain three-row TextField components. */ .v-table-components-inside .v-table-cell-content { height: 54px; }
The table will look as shown in Figure 5.50, “Components in a Table”.
As the items in a Table
are not indexed, iterating
over the items has to be done using an iterator. The
getItemIds()
method of the
Container
interface of Table
returns a Collection
of item identifiers over which
you can iterate using an Iterator
. For an example
about iterating over a Table
, please see Section 9.4, “Collecting Items in Containers”. Notice that you may not modify the
Table
during iteration, that is, add or remove
items. Changing the data is allowed.
A table can be filtered if its container data source implements the
Filterable
interface, as the default
IndexedContainer
does. See Section 9.4.4, “Filterable
Containers”.
Normally, a Table
simply displays the items and
their fields as text. If you want to allow the user to edit the values,
you can either put them inside components as we did above, or you can
simply call setEditable(true)
and the cells are
automatically turned into editable fields.
Let us begin with a regular table with a some columns with usual Java
types, namely a Date
,
Boolean
, and a String
.
// Create a table. It is by default not editable. final Table table = new Table(); // Define the names and data types of columns. table.addContainerProperty("Date", Date.class, null); table.addContainerProperty("Work", Boolean.class, null); table.addContainerProperty("Comments", String.class, null); // Add a few items in the table. for (int i=0; i<100; i++) { Calendar calendar = new GregorianCalendar(2008,0,1); calendar.add(Calendar.DAY_OF_YEAR, i); // Create the table row. table.addItem(new Object[] {calendar.getTime(), new Boolean(false), ""}, new Integer(i)); // Item identifier } table.setPageLength(8); layout.addComponent(table);
You could put the table in editable mode right away if you need to. We'll
continue the example by adding a mechanism to switch the
Table
from and to the editable mode.
final CheckBox switchEditable = new CheckBox("Editable"); switchEditable.addListener(new Property.ValueChangeListener() { public void valueChange(ValueChangeEvent event) { table.setEditable(((Boolean)event.getProperty() .getValue()).booleanValue()); } }); switchEditable.setImmediate(true); layout.addComponent(switchEditable);
Now, when you check to checkbox, the components in the table turn into editable fields, as shown in Figure 5.51, “A Table in Normal and Editable Mode”.
The field components that allow editing the values of particular types
in a table are defined in a field factory that implements the
TableFieldFactory
interface. The default
implementation is DefaultFieldFactory
, which
offers the following crude mappings:
Table 5.2. Type to Field Mappings in DefaultFieldFactory
Property Type | Mapped to Field Class |
---|---|
Date | A DateField . |
Boolean | A CheckBox . |
Item | A Form . The fields of the
form are automatically created from the item's properties
using a FormFieldFactory . The
normal use for this property type is inside a
Form and is less useful inside a
Table . |
other | A TextField . The text field
manages conversions from the basic types, if
possible. |
Field factories are covered with more detail in Section 5.19.2, “Binding Form to Data”. You could just implement the
TableFieldFactory
interface, but we recommend that
you extend the DefaultFieldFactory
according to
your needs. In the default implementation, the mappings are defined in the
createFieldByPropertyType()
method (you might
want to look at the source code) both for tables and forms.
In the editable mode, the editor fields can have focus. Pressing Tab moves the focus to next column or, at the last column, to the first column of the next item. Respectively, pressing Shift+Tab moves the focus backward. If the focus is in the last column of the last visible item, the pressing Tab moves the focus outside the table. Moving backward from the first column of the first item moves the focus to the table itself. Some updates to the table, such as changing the headers or footers or regenerating a column, can move the focus from an editor component to the table itself.
The default behaviour may be undesirable in many cases. For example, the focus also goes through any read-only editor fields and can move out of the table inappropriately. You can provide better navigation is to use event handler for shortcut keys such as Tab, Arrow Up, Arrow Down, and Enter.
// Keyboard navigation class KbdHandler implements Handler { Action tab_next = new ShortcutAction("Tab", ShortcutAction.KeyCode.TAB, null); Action tab_prev = new ShortcutAction("Shift+Tab", ShortcutAction.KeyCode.TAB, new int[] {ShortcutAction.ModifierKey.SHIFT}); Action cur_down = new ShortcutAction("Down", ShortcutAction.KeyCode.ARROW_DOWN, null); Action cur_up = new ShortcutAction("Up", ShortcutAction.KeyCode.ARROW_UP, null); Action enter = new ShortcutAction("Enter", ShortcutAction.KeyCode.ENTER, null); public Action[] getActions(Object target, Object sender) { return new Action[] {tab_next, tab_prev, cur_down, cur_up, enter}; } public void handleAction(Action action, Object sender, Object target) { if (target instanceof TextField) { // Move according to keypress int itemid = (Integer) ((TextField) target).getData(); if (action == tab_next || action == cur_down) itemid++; else if (action == tab_prev || action == cur_up) itemid--; // On enter, just stay where you were. If we did // not catch the enter action, the focus would be // moved to wrong place. if (itemid >= 0 && itemid < table.size()) { TextField newTF = valueFields.get(itemid); if (newTF != null) newTF.focus(); } } } } // Panel that handles keyboard navigation Panel navigator = new Panel(); navigator.addStyleName(Reindeer.PANEL_LIGHT); navigator.addComponent(table); navigator.addActionHandler(new KbdHandler());
The main issue in implementing keyboard navigation in an editable
table is that the editor fields do not know the table they are in. To
find the parent table, you can either look up in the component
container hierarchy or simply store a reference to the table with
setData()
in the field component. The other
issue is that you can not acquire a reference to an editor field from
the Table
component. One solution is to use
some external collection, such as a HashMap
, to
map item IDs to the editor fields.
// Can't access the editable components from the table so // must store the information final HashMap<Integer,TextField> valueFields = new HashMap<Integer,TextField>();
The map has to be filled in a
TableFieldFactory
, such as in the
following. You also need to set the reference to the table there and
you can also set the initial focus there.
table.setTableFieldFactory(new TableFieldFactory () { public Field createField(Container container, Object itemId, Object propertyId, Component uiContext) { TextField field = new TextField((String) propertyId); // User can only edit the numeric column if ("Source of Fear".equals(propertyId)) field.setReadOnly(true); else { // The numeric column // The field needs to know the item it is in field.setData(itemId); // Remember the field valueFields.put((Integer) itemId, field); // Focus the first editable value if (((Integer)itemId) == 0) field.focus(); } return field; } });
The issues are complicated by the fact that the editor fields are not generated for the entire table, but only for a cache window that includes the visible items and some items above and below it. For example, if the beginning of a big scrollable table is visible, the editor component for the last item does not exist. This issue is relevant mostly if you want to have wrap-around navigation that jumps from the last to first item and vice versa.
Table
supports both column headers and footers; the
headers are enabled by default.
The table header displays the column headers at the top of the
table. You can use the column headers to reorder or resize the
columns, as described earlier. By default, the header of a column is
the property ID of the column, unless given explicitly with
setColumnHeader()
.
// Define the properties table.addContainerProperty("lastname", String.class, null); table.addContainerProperty("born", Integer.class, null); table.addContainerProperty("died", Integer.class, null); // Set nicer header names table.setColumnHeader("lastname", "Name"); table.setColumnHeader("born", "Born"); table.setColumnHeader("died", "Died");
The text of the column headers and the visibility of the header
depends on the column header mode. The header is
visible by default, but you can disable it with
setColumnHeaderMode(Table.COLUMN_HEADER_MODE_HIDDEN)
.
The table footer can be useful for displaying sums or averages of
values in a column, and so on. The footer is not visible by default;
you can enable it with
setFooterVisible(true)
. Unlike in the header,
the column headers are empty by default. You can set their value with
setColumnFooter()
. The columns are identified
by their property ID.
The following example shows how to calculate average of the values in a column:
// Have a table with a numeric column Table table = new Table("Custom Table Footer"); table.addContainerProperty("Name", String.class, null); table.addContainerProperty("Died At Age", Integer.class, null); // Insert some data Object people[][] = {{"Galileo", 77}, {"Monnier", 83}, {"Vaisala", 79}, {"Oterma", 86}}; for (int i=0; i<people.length; i++) table.addItem(people[i], new Integer(i)); // Calculate the average of the numeric column double avgAge = 0; for (int i=0; i<people.length; i++) avgAge += (Integer) people[i][1]; avgAge /= people.length; // Set the footers table.setFooterVisible(true); table.setColumnFooter("Name", "Average"); table.setColumnFooter("Died At Age", String.valueOf(avgAge)); // Adjust the table height a bit table.setPageLength(table.size());
The resulting table is shown in Figure 5.52, “A Table with a Footer”.
Normally, when the user clicks a column header, the table will be
sorted by the column, assuming that the data source is
Sortable
and sorting is not disabled. In some
cases, you might want some other functionality when the user clicks
the column header, such as selecting the column in some way.
Clicks in the header cause a HeaderClickEvent
,
which you can handle with a
Table.HeaderClickListener
. Click events on the
table header (and footer) are, like button clicks, sent immediately to
server, so there is no need to set
setImmediate()
.
// Handle the header clicks table.addListener(new Table.HeaderClickListener() { public void headerClick(HeaderClickEvent event) { String column = (String) event.getPropertyId(); getWindow().showNotification("Clicked " + column + "with " + event.getButtonName()); } }); // Disable the default sorting behavior table.setSortDisabled(true);
Setting a click handler does not automatically disable the sorting
behavior of the header; you need to disable it explicitly with
setSortDisabled(true)
. Header click events
are not sent when the user clicks the column resize handlers to drag
them.
The HeaderClickEvent
object provides the
identity of the clicked column with
getPropertyId()
. The
getButton()
reports the mouse button with
which the click was made: BUTTON_LEFT
,
BUTTON_RIGHT
, or
BUTTON_MIDDLE
. The
getButtonName()
a human-readable button name
in English: "left
",
"right
", or
"middle
". The
isShiftKey()
,
isCtrlKey()
, etc., methods indicate if the
Shift, Ctrl, Alt or
other modifier keys were pressed during the click.
Clicks in the footer cause a FooterClickEvent
,
which you can handle with a
Table.FooterClickListener
. Footers do not have
any default click behavior, like the sorting in the header. Otherwise,
handling clicks in the footer is equivalent to handling clicks in the
header.
You might want to have a column that has values calculated from other
columns. Or you might want to format table columns in some way, for
example if you have columns that display currencies. The
ColumnGenerator
interface allows defining custom
generators for such columns.
You add new generated columns to a Table
with
addGeneratedColumn()
. It takes the column
identifier as its parameters. Usually you want to have a more
user-friendly and possibly internationalized column header. You can set
the header and a possible icon by calling
addContainerProperty()
before adding the generated column.
// Define table columns. table.addContainerProperty( "date", Date.class, null, "Date", null, null); table.addContainerProperty( "quantity", Double.class, null, "Quantity (l)", null, null); table.addContainerProperty( "price", Double.class, null, "Price (e/l)", null, null); table.addContainerProperty( "total", Double.class, null, "Total (e)", null, null); // Define the generated columns and their generators. table.addGeneratedColumn("date", new DateColumnGenerator()); table.addGeneratedColumn("quantity", new ValueColumnGenerator("%.2f l")); table.addGeneratedColumn("price", new PriceColumnGenerator()); table.addGeneratedColumn("total", new ValueColumnGenerator("%.2f e"));
Notice that the addGeneratedColumn()
always
places the generated columns as the last column, even if you defined some
other order previously. You will have to set the proper order with
setVisibleColumns()
.
table.setVisibleColumns(new Object[] {"date", "quantity", "price", "total"});
The generators are objects that implement the
Table.ColumnGenerator
interface and its
generateCell()
method. The method gets the
identity of the item and column as its parameters, in addition to the
table object. It has to return a component object.
The following example defines a generator for formatting
Double
valued fields according to a format string
(as in java.util.Formatter
).
/** Formats the value in a column containing Double objects. */ class ValueColumnGenerator implements Table.ColumnGenerator { String format; /* Format string for the Double values. */ /** * Creates double value column formatter with the given * format string. */ public ValueColumnGenerator(String format) { this.format = format; } /** * Generates the cell containing the Double value. * The column is irrelevant in this use case. */ public Component generateCell(Table source, Object itemId, Object columnId) { // Get the object stored in the cell as a property Property prop = source.getItem(itemId).getItemProperty(columnId); if (prop.getType().equals(Double.class)) { Label label = new Label(String.format(format, new Object[] { (Double) prop.getValue() })); // Set styles for the column: one indicating that it's // a value and a more specific one with the column // name in it. This assumes that the column name // is proper for CSS. label.addStyleName("column-type-value"); label.addStyleName("column-" + (String) columnId); return label; } return null; } }
The generator is called for all the visible (or more accurately cached) items in a table. If the user scrolls the table to another position in the table, the columns of the new visible rows are generated dynamically. The columns in the visible (cached) rows are also generated always when an item has a value change. It is therefore usually safe to calculate the value of generated cells from the values of different rows (items).
When you set a table as editable
, regular fields
will change to editing fields. When the user changes the values in the
fields, the generated columns will be updated automatically. Putting a
table with generated columns in editable mode has a few quirks. The
editable mode of Table
does not affect generated
columns. You have two alternatives: either you generate the editing fields
in the generator or, in case of formatter generators, remove the generator
in the editable mode. The example below uses the latter approach.
// Have a check box that allows the user // to make the quantity and total columns editable. final CheckBox editable = new CheckBox( "Edit the input values - calculated columns are regenerated"); editable.setImmediate(true); editable.addListener(new ClickListener() { public void buttonClick(ClickEvent event) { table.setEditable(editable.booleanValue()); // The columns may not be generated when we want to // have them editable. if (editable.booleanValue()) { table.removeGeneratedColumn("quantity"); table.removeGeneratedColumn("total"); } else { // Not editable // Show the formatted values. table.addGeneratedColumn("quantity", new ValueColumnGenerator("%.2f l")); table.addGeneratedColumn("total", new ValueColumnGenerator("%.2f e")); } // The visible columns are affected by removal // and addition of generated columns so we have // to redefine them. table.setVisibleColumns(new Object[] {"date", "quantity", "price", "total", "consumption", "dailycost"}); } });
You will also have to set the editing fields in
immediate
mode to have the update occur immediately
when an edit field loses the focus. You can set the fields in
immediate
mode with the a custom
TableFieldFactory
, such as the one given below,
that just extends the default implementation to set the mode:
public class ImmediateFieldFactory extends DefaultFieldFactory { public Field createField(Container container, Object itemId, Object propertyId, Component uiContext) { // Let the DefaultFieldFactory create the fields... Field field = super.createField(container, itemId, propertyId, uiContext); // ...and just set them as immediate. ((AbstractField)field).setImmediate(true); return field; } } ... table.setTableFieldFactory(new ImmediateFieldFactory());
If you generate the editing fields with the column generator, you avoid having to use such a field factory, but of course have to generate the fields for both normal and editable modes.
Figure 5.53, “Table with Generated Columns in Normal and Editable Mode” shows a table with columns calculated (blue) and simply formatted (black) with column generators.
The displayed values of properties shown in a table are normally formatted
using the toString()
method of each
property. Customizing the format of a column can be done in several ways:
ColumnGenerator
to generate a
second column that is formatted. The original column needs to be set
invisible. See Section 5.14.5, “Generated Table Columns”.PropertyFormatter
as a proxy
between the table and the data property. This also normally requires using
an mediate container in the table.formatPropertyValue()
in
Table
.
As using a PropertyFormatter
is generally much more
awkward than overriding the
formatPropertyValue()
, its use is not described
here.
You can override formatPropertyValue()
as is done
in the following example:
// Create a table that overrides the default // property (column) format final Table table = new Table("Formatted Table") { @Override protected String formatPropertyValue(Object rowId, Object colId, Property property) { // Format by property type if (property.getType() == Date.class) { SimpleDateFormat df = new SimpleDateFormat("yyyy-MM-dd hh:mm:ss"); return df.format((Date)property.getValue()); } return super.formatPropertyValue(rowId, colId, property); } }; // The table has some columns table.addContainerProperty("Time", Date.class, null); ... Fill the table with data ...
You can also distinguish between columns by the
colId
parameter, which is the property ID of the
column. DecimalFormat
is useful for formatting
decimal values.
... in formatPropertyValue() ... } else if ("Value".equals(pid)) { // Format a decimal value for a specific locale DecimalFormat df = new DecimalFormat("#.00", new DecimalFormatSymbols(locale)); return df.format((Double) property.getValue()); } ... table.addContainerProperty("Value", Double.class, null);
A table with the formatted date and decimal value columns is shown in Figure 5.54, “Formatted Table Columns”.
You can use CSS for further styling of table rows, columns, and individual
cells by using a CellStyleGenerator
. It is
described in Section 5.14.7, “CSS Style Rules”.
Styling the overall style of a Table
can be done
with the following CSS rules.
.v-table {} .v-table-header-wrap {} .v-table-header {} .v-table-header-cell {} .v-table-resizer {} /* Column resizer handle. */ .v-table-caption-container {} .v-table-body {} .v-table-row-spacer {} .v-table-table {} .v-table-row {} .v-table-cell-content {}
Notice that some of the widths and heights in a table are calculated dynamically and can not be set in CSS.
The Table.CellStyleGenerator
interface allows
you to set the CSS style for each individual cell in a table. You need
to implement the getStyle()
, which gets the
row (item) and column (property) identifiers as parameters and can
return a style name for the cell. The returned style name will be
concatenated to prefix "v-table-cell-content-
".
The getStyle()
is called also for each row,
so that the propertyId
parameter is
null
. This allows setting a row style.
Alternatively, you
can use a Table.ColumnGenerator
(see Section 5.14.5, “Generated Table Columns”) to generate the actual UI
components of the cells and add style names to them.
Table table = new Table("Table with Cell Styles"); table.addStyleName("checkerboard"); // Add some columns in the table. In this example, the property // IDs of the container are integers so we can determine the // column number easily. table.addContainerProperty("0", String.class, null, "", null, null); for (int i=0; i<8; i++) table.addContainerProperty(""+(i+1), String.class, null, String.valueOf((char) (65+i)), null, null); // Add some items in the table. table.addItem(new Object[]{ "1", "R", "N", "B", "Q", "K", "B", "N", "R"}, new Integer(0)); table.addItem(new Object[]{ "2", "P", "P", "P", "P", "P", "P", "P", "P"}, new Integer(1)); for (int i=2; i<6; i++) table.addItem(new Object[]{String.valueOf(i+1), "", "", "", "", "", "", "", ""}, new Integer(i)); table.addItem(new Object[]{ "7", "P", "P", "P", "P", "P", "P", "P", "P"}, new Integer(6)); table.addItem(new Object[]{ "8", "R", "N", "B", "Q", "K", "B", "N", "R"}, new Integer(7)); table.setPageLength(8); // Set cell style generator table.setCellStyleGenerator(new Table.CellStyleGenerator() { public String getStyle(Object itemId, Object propertyId) { // Row style setting, not relevant in this example. if (propertyId == null) return "green"; // Will not actually be visible int row = ((Integer)itemId).intValue(); int col = Integer.parseInt((String)propertyId); // The first column. if (col == 0) return "rowheader"; // Other cells. if ((row+col)%2 == 0) return "black"; else return "white"; } });
You can then style the cells, for example, as follows:
/* Center the text in header. */ .v-table-header-cell { text-align: center; } /* Basic style for all cells. */ .v-table-checkerboard .v-table-cell-content { text-align: center; vertical-align: middle; padding-top: 12px; width: 20px; height: 28px; } /* Style specifically for the row header cells. */ .v-table-cell-content-rowheader { background: #E7EDF3 url(../default/table/img/header-bg.png) repeat-x scroll 0 0; } /* Style specifically for the "white" cells. */ .v-table-cell-content-white { background: white; color: black; } /* Style specifically for the "black" cells. */ .v-table-cell-content-black { background: black; color: white; }
The table will look as shown in Figure 5.55, “Cell Style Generator for a Table”.