Develop Reference

Recreate Jtable or remove all rows and add new rows?

I have a JTable and a DefaultTableModel, I want to delete all the rows of JTable every 3 secondes and add new rows to the JTable. Considering the performance, What's the best choice to do that, recreate Jtable or remove all rows and add rows?

What's the best choice to do that, renew Jtable or remove all rows and addrows?
If you recreate the JTable, all the renderers and editors will be recreated for the table.
If you change the TableModel, then the table will need to recreate the TableColumnModel, and all the TableColumns based on the new TableModel.
There will be many more objects that need to be created to support the JTable/TableModel relationship.
Removing rows and adding rows, will just cause the table to repaint the new data using all the current renderers. I would say this is more efficient. And instead of adding rows one at a time, you should add all the rows at once

You can create a tablemodel that digs into your data object. As soon as you change the underlying data object, you call fireTableDataChanged().
From the JavaDoc for AbstractTableModel#fireTableDataChanged():
Notifies all listeners that all cell values in the table's rows may have changed. The number of rows may also have changed and the JTable should redraw the table from scratch. The structure of the table (as in the order of the columns) is assumed to be the same.

Javax Swing JTable::getModel vs JTable::getColumnModel

Hello I have been working with javax swing and I came across a weird problem and got me questioning.
I can for example:
JTable table = new JTable();
// Indeed, 2 different objects:
// The TableModel (which, i think is supposed to contain rows and columns?
DefaultTableModel dtm = (DefaultTableModel) table.getModel();
// And the column model, supposed to define the columns of a table?
TableColumnModel tcm = table.getColumnModel();
// I can add columns to my table in two different manners
dtm.addColumn("A Column");
// or
TableColumn column = new TableColumn();
column.setHeaderValue("Another column");
column.setWidth(120);
column.setMinWidth(120);
column.setMaxWidth(120);
tcm.addColumn(column);
// And notice that both commands will add a column in the table
// So our table model should now have 2 columns.
// But does it?
System.out.println(dtm.getColumnCount()); // outputs 1;
System.out.println(tcm.getColumnCount()); // outputs 2;
System.out.println(table.getColumnCount()); // outputs 2;
// The visual shows 2 columns, but the model has only 1.
From that I can tell JTable uses tableColumnModel and tableColumnModel gets all the columns added into tableModel, but, when I add a column to the TableModel it gets added to the table, but the tableModel remains outdated.
Now, the problem is: it's really interesting to add a column via columnModel because I can define the sizes, layout, editable options there, but in this way I cannot add any data to it from the tableModel, since that column doesn't appear on the tableModel.
Any thoughts on this?

The TableModel is used to contain data. The data is accessible in row/columns.
The TableColumnModel is used by JTable to control the View of the data. That is it controls the columns that are displayed in the JTable. You can also reorder the columns to display the data in a different order.
...but in this way I cannot add any data to it from the tableModel, since that column doesn't appear on the tableModel
That is correct. The purpose of the TableColumnModel is to simply customize the view, not manipulate data.
Maybe you have an application that contains many columns of data, but access to specific columns is limited by "security level". In this case the data is always stored in the TableModel, but you need to change the view to control which columns of data are visible. So you can remove/add columns from the TableColumnModel.
When you add a column to the TableModel, the JTable gets notified and it recreates all the TableColumns for you. This can be a good or bad thing because when the TableColumnModel is recreated you lose any custom renderers and editor that you may have added to a TableColumn. You can prevent this from happening buy using:
table.setAutoCreateColumnsFromModel( false );
Now the TableColumnModel will not be updated and it is your responsibility to manually create and add the TableColumn to the TableColumnModel.
But in general you:
add/change data through the TableModel.
change the view through the TableColumnModel.

Which editing and removing row method is better for JTable?

I wanna ask that which method to edit a JTable is better than the other?
use setValueAt() function.
Make a new model of JTable and feed its object to JTable.
And also, how to remove a Row from JTable without using method 2 described above?

You would use the setValueAt method when you're changing one value at a time.
You would make a new TableModel when you're changing 25% or more of the rows at one time.
If you use a DefaultTableModel as your TableModel, you can use the addRow or setValueAt methods to add a row and change a row value. You can use the removeRow method to remove a row.
Since the TableModel is associated with the JTable, any changes to the TableModel are reflected in the JTable. Any changes to the JTable by the user are reflected in the TableModel.

The best way to set rows and headers separately after JTable creation?

It means I need setData and setHeaders functions for JTable

Like Gilbert says, there is no method of the AbstractTableModel that will allow you to set the values of the headers after the JTable is up, since it is not a very common requirement. Still there is a simple work-around to that. I can not say if it is the best way to do it, but it will get you there...
columnNumber is the number of the column you want to change and newHeaderString is the new String you want to use.
jTable1.getColumnModel().getColumn(columnNumber).setHeaderValue(newHeaderString);
jScrollPane1.setViewportView(jTable1);
Since there is a getTableHeader() method, you could call it if you need to further modify your table header properties.
As for setting new data in any row, use jTable1.setValueAt(newObject, row, col);

if you use a TableModel, you can change the data contents, but not the column headers.
TableModel model = new DefaultTableModel(rowData, columnNames);
JTable table = new JTable(model)
rowData and columnNames have to be defined before you create the table.
The TableModel interface has a setValueAt method. The TableModel interface has no method for setting column names.

Dynamically changing the column header text in JTable

I have a table with 3 columns which have the following values in the headers: 'No.', 'X [mm]', 'Y [mm]'. This table contain the coordinates of points in millimeters. I have a checkbox on checking of which the table should repopulate to show the coordinates in inches. Moreover, the column header values should be: 'No.', 'X [in]', 'Y [in]'.
In short I want to dynamically change the header text of the table.
In detail:
The table is a subclass of JTable. Moreover, a subclass of 'DefaultTableModel' has been set as the model for the table. I have provided the header values in the constructer of the datamodel subclass.
Any idea? My application is compatible with only jdk v1.4 so it would be good if the solution is compatible with the verion :)

You can update the TableColumnModel directly:
JTableHeader th = table.getTableHeader();
TableColumnModel tcm = th.getColumnModel();
TableColumn tc = tcm.getColumn(0);
tc.setHeaderValue( "???" );
th.repaint();

If you have column number use that code
jtable.getColumnModel().getColumn(5).setHeaderValue("newHeader");

I can't test here but familiar that this method '[DefaultTableModel.setColumnIdentifiers(...)][1]' should do what you want.
Basically, you run 'DefaultTableModel.getColumnCount()' to find out how many column (unless you already know). Then you run 'DefaultTableModel.getColumnName(int ColumnIndex)' to get the name of each, change it the way you want and put it in an array. After thatn, you set them back using 'DefaultTableModel.setColumnIdentifiers(...)'.
Hope this helps.