he latest version of Java, J2SE 5.0 (version 1.5 of the JDK) is the first to deliver the concept of disconnected RowSet objects to the Java language. The FilteredRowset interface extends from the WebRowSet interface, which in turn extends from the javax.sql.Rowset interface. As you will see, the FilteredRowSet lets one narrow down the number of rows in a disconnected object based on filtering logic you provide without requiring an ongoing connection to your database. In this article, I'll introduce you to FilteredRowSet objects.

Where Oh WHERE Clause?
If you were using simple JDBC, you could achieve the same thing using a WHERE clause in a query using a JdbcRowset object. But JdbcRowset objects require a connection to the database, whereas RowSet objects (a superset of FilteredRowSet objects) do not (i.e., they are disconnected). A WHERE clause requires a connection to the database to filter your database data. With the FilteredRowSet, you can retrieve your data from the database and then disconnect. As you will see, the FilteredRowSet uses Predicate objects to retrieve data from RowSet objects without a database connection using WHERE-like logic.

A Real World Example
I've decided to build a fun example to help facilitate understanding. This application will retrieve NBA (National Basketball Association) player statistics for the 2003-04 season. Don't worry, you don't need to know much about basketball for this study. You'll build a database table that stores player data including games played (G), field goals made (FG), free throws made (FT), total points scored (P), and average points per game (PG).

To do this, you'll need a database and, of course, a database table. You can choose any database you prefer but it must support FilteredRowSet objects, which are an offering of Java 1.5. I am going to use IBM's DB2 8.1, which you can download as a trial from http://www-306.ibm.com/software/data/db2/udb/v8/.

I'll start by creating a database using the DB2 command line processor:

db2 => create database balldb

Next, I'll connect to the database with my user name (db2admin) and password (db2admin).

db2=> connect to balldb user db2admin using db2admin

Next, I'll create a table (named STATS) to house all the statistical data mentioned above. The following statement creates the table:

db2=> create table stats (firstname varchar(40) not null, lastname varchar(40) 
   not null, team varchar(40) not null, G int not null, FG int not null, FT int not null, P int not null,
 PG decimal(3,1) not null)

I need some data in my table. In the real world, this table would likely have data for all the players in the NBA. For the purposes of this example, I've included only 10 players in the table—namely the players with the highest average points per game during the 2003-2004 regular season:

Table 1. The NBA Table Data.

FIRSTNAME	LAST NAME	TEAM	G	FG	FT	P	PG
Tracy	McGrady	Orlando Magic	67	653	398	1,878	28.0
Predrag	Stojakovic	Sacremento Kings	81	665	394	1,964	24.2
Kevin	Garnett	Minnesota Timberwolves	82	804	368	1,987	24.2
Kobe	Bryant	Los Angeles Lakers	65	516	454	1,557	24.0
Paul	Pierce	Boston Celtics	80	602	517	1,836	23.0
Baron	Davis	New Orleans Hornets	67	554	237	1,532	22.9
Vince	Carter	Toronto Raptors	73	608	336	1,645	22.5
Tim	Duncan	San Antonio Spurs	69	592	352	1,538	22.3
Dirk	Nowitzki	Dallas Mavericks	77	605	371	1,680	21.8
Michael	Redd	Milwaukee Bucks	82	633	383	1,776	21.7

The following statement inserts the first row into the database.

insert into stats values ('Tracy','McGrady','Orlando Magic',67,653,398,1878,28.0)

You can add each of the remaining nine rows with appropriate data using the same process, until the table is complete.