PFJDBC driver it's a wrapper for any real JDBC driver which enables developers and DBAs in early stages of deployment and during integration tests to take a deep look at the interaction between the program and the database.

  In other words, it performs an analysis for all the JDBC code that exists in a Java program, capturing both static and dynamic behaviour of the code. To achieve this goal, it has to substitute to the real JDBC driver, intercept all the calls and print its measurments and hints into some output stream.

  This driver was created to be easy to use by all people involved into a project. You will not see in the configuration files long an incomprehensible classnames, the output can be suppressed in a matter of seconds by just changing one of the configuration parameters; this lets the developer or DBA to concentrate on the solving database specific problems not on the configuration issues.

  One of the benefits this driver adds is that no additional code is required in your program; it eliminates the need for debug statements mixed within your code. You can forget about pieces of code like:

long startTime = System.currentTimeMillis();
 ... //do whatever you do with SQL
long intvl = System.currentTimeMillis() - startTime;
LOG.log(Level.FINEST, "SQL took: " + intvl + " ms");

  So, you can use this driver to reach 4 main objectives:

1. To measure the performance of your SQL phrases, recordsets, transactions, etc. You can then take the most frequent ones and improve the execution time.
2. To validate the correctness of your program, from the database point of view.
3. To detect possible performance problems (like cache and bottlenecks at database side)
4. To improve code (accessors, use of not recommended syntax, etc)