How do I use use data from a Resultset in a JTable?

Question

I am doing a school project and I am having trouble with storing the data from a resultset in a JTable. Previously I had used DButils but now I am wondering if there is a way to do the same thing without having to used external class files or if it is easier to use DButils.

The data is coming from only one table and all that needs to happen is the data must be displayed in the JTable.

I would post my code here but I have looked and the only tutorials I could find were ones on how to populate a JTable using and Object [][]. I am using JDBC to create the connection.

Thanks in advance.

Answer 1

Of course there's a way: iterate through the ResultSet and add what you find to the Object [][] array that gets passed to the JTable. There's one row in the 2D array for each row in the ResultSet; the columns are the values.

The problem you'll have is that you won't know how many rows came back without iterating through it. That's why loading it into a Map<String, Object> might be a better idea.

Here's an example showing how to do it. You'll find that method (and more) at my answer to this question:

java sql connections via class

public static List<Map<String, Object>> map(ResultSet rs) throws SQLException {
    List<Map<String, Object>> results = new ArrayList<Map<String, Object>>();
    try {
        if (rs != null) {
            ResultSetMetaData meta = rs.getMetaData();
            int numColumns = meta.getColumnCount();
            while (rs.next()) {
                Map<String, Object> row = new HashMap<String, Object>();
                for (int i = 1; i <= numColumns; ++i) {
                    String name = meta.getColumnName(i);
                    Object value = rs.getObject(i);
                    row.put(name, value);
                }
                results.add(row);
            }
        }
    } finally {
        close(rs);
    }
    return results;
}

Answer 2

Well this will require several steps.

I will explain my way, which is good for very large sets, but a little complicated if you only want to show a few lines. Still I'm sure it will help you. This method will load the required records on the fly, not the whole set before hand. It creates the illusion of having the whole set, but with out having to wait for a lengthy load.

1) Ok, first, let's assume that we have a nice JFrame that you can display, to start with. So first I will add a JScrollPane, and inside it I will add a JTable. Run it and make sure you have a nice window with an empty JTable inside scroll bars.

2) So next you need a data source for the JTable. Since a JTable is a very generic component not made specifically for SQL resultSets, it requires a data source that implements javax.swing.table.AbstractTableModel which has nothing to do with SQL. So we will now create a TableModelClass which will implement AbstractTableModel, then we will add this to the JTable and it will start working. Of course, the trick is to implement all of AbstractTableModel's methods that get data by using our SQL result set, and this is up to you. From here is my suggestion ->

3) Since this will be dynamic, we will not need to load all the data before hand, but we need an initial set to display. I will have a Object [][] of a fixed size, lets say 200 - 300 rows. So I will initially execute the SQL and fill the array with the buffer size of 200-300 rows. How much to cache will depend on 2 things: 1 It shall be enough to get all the data for the current display size of the JTable, and 2, it should be small enough so that as we scroll and get subsequent caches it executes very fast.

4) Now let's begin implementing all AbstractTableModel's interface methods.

5) First we look at the initial result set and report the number of columns. Just add a class variable, set the column count and return it using: public int getColumnCount( ). This will not change from now.

6) Also looking at the result set metadata, make a list variable in the class and add the column names returned in the meta data. Using this list return the column names in "getColumnName( int col )". Of course, the col index is the column position in the result set.

7) Now lets do "int getRowCount( )". Inside the TableModelClass keep a variable to contain the rowCount and return it in this method. TIP: Don’t worry for now, set it to a fixed large number like 65000, this will let scroll as you dynamically load the records. Once we hit the end we will set the number to the real value and the scroll pane will adjust to the correct proportions. Trust me, it works ok.

8) Now comes the fun part. As the JTable presents the first "page" of the table and as the user scrolls it will begin calling "getValueAt( int row, int col )". This will map directly to our Object[][], but since we only have a cache, and not the whole table, as the user scrolls down we will need to fetch more data. I do this:

 public Object getValueAt( int row, int col )
    {
        // load failed before, no more trying...
        if( loadExceptionOccur || ( row >= visualTableSize ) ) return( "" );

        // check if requested row is OUT of cache …
        try{             
        if( 
            // less than cache lower limit...
            ( ( row < startRow ) 
            || 
            // Beyond cache upper limit...
            ( row >= startRow + tableDataCache.size()) ) 
                        // Stop unnecessary loading, because of Jtable readjusting 
                        // its visual table size and redrawing the entire table.
                        && !tableRedraw

        // yes, get new cache...
        ){ 
            load( row );   // <- below is code
        }

            // now we now the row is in cache, so ->

            // verify requested cell in cache, or beyond data rows, 
            if( 
                    // greater than lower limit
                    ( row >= startRow ) 
                    && 
                    // less than upper limit...
                    ( row < ( startRow + tableDataCache.size() ) ) 
            ){
                    tableRedraw = false;

             // just get the data from the cache. tableDataCache is just your Object[][] array…
                    Object cellValue = ( ( recordClass ) tableDataCache.get( row-startRow ) ).getValueAt( col );


                    return ( cellValue );
            }
            else{
                    // just show as blank
                    return( "" );            
            }                
        }
        catch( Exception error  ){



    …

In case of a cache miss you need to reload a cache of data. I will normally load some rows before the requested row and some beyond, at least for a JTable page size, so that we only go once to the db to render a screen. The bigger the cache the more scrolling before loading, but the larger the time it takes to load a cache. If you fine tune it, the cache processing might be almost unnoticeable.

Here is the implementation of "load":

public void load( int rowIndex )
    throws KExceptionClass
    {
        // calculate start of new cache, if not enough rows for top half of cache 
        // then start from 0
        int halfCache = cacheSize / 2 ;
        int DBStartRow = 0;       
        if( rowIndex > halfCache ) DBStartRow = rowIndex - halfCache;

        //Do query to DB
        try{

            SQLP.load( DBStartRow, cacheSize );   // <- using jdbc load from DbsartRow as many rows as cacheSize. Some sample SQL paging code below ->

        }catch( Exception loadError ){

           // if the database fails or something do this, so you don’t get a billion errors for each cell. ->

            //set load failed flag, kill window                        
            loadExceptionOccur = true;         
            visualTableSize = 0;
            tableDataCache = new ArrayList< recordClass >();        
            fireTableDataChanged();  // clear the Jtable 

            // log error
            log.log( this, KMetaUtilsClass.getStackTrace( loadError ) );
            // show error message
            throw new KExceptionClass( "Could not load table data! " , loadError );

        }

        //Load rows into the cache list.
        //Key field values are in the cache list as the last field in each record.
        tableDataCache.clear();       // the Object [][], wrapped in class

        while( SQLPreprocessor.nextRowValue() ) { 

            SQL.fetch( record );             //<- get JDBC rows to table cache
            tableDataCache.add( record );    // this uses my library, change to JDBC or what ever you use to access SQL
        }

        log.log( this, "cacheList size = " + tableDataCache.size());

        //---------

        if( 
            // Last requested row number        
            ( DBStartRow + cacheSize )  > 
            // Last replied row number
            ( SQLPreprocessor.getloadedStartRowIndex() + SQLPreprocessor.getloadedRowCount() ) 
            ){ 
                // It is the end of table.
                // The visual table is readjusted accordingly. 
                visualTableSize = SQLPreprocessor.getloadedStartRowIndex() + SQLPreprocessor.getloadedRowCount();   
                fireTableDataChanged(); 
                tableRedraw = true;
            }

        startRow = SQLPreprocessor.getloadedStartRowIndex();                        


        log.log( this, "visualTableSize = " + visualTableSize );  

    }

Ok this will dynamically load the data in small caches which will give the impression of having the whole set. If the user scrolls to the middle or all the way to the end, the JTable will ask only for the data need to display not all the rows as it moves, so, if you have a 10K row table, but the JTable is only 20 rows high, a scroll to the end will only take 40 - 50 rows to load. Pretty nice. Your users will be impressed.

Now the thing is that the load assumes that you have a SQL cursor that moves forward and backwards by row number. This simple thing is quite a challenge in SQL. For Oracle check : http://www.oracle.com/technetwork/issue-archive/2006/06-sep/o56asktom-086197.html

Ok, hope that helps.--