Matt Casters has added SQLite support to Pentaho’s Kettle ETL tool in the latest development release. I’ve tested it under Windows using JRE 1.5.0_09 and it worked fine but having upgraded to JRE 1.5.0._10 I’m now getting “out of memory” errors, appears to be a problem with the “pure java” jdbc driver and not with Kettle itself.
Problem persists with JRE 1.5.0._11, but the bug appears to be in how the driver responds to window’s drive letters, so a work-around is to use unix-style folder addressing. For example, if your database is “c:\kettle\data\mydatabase.db” and Kettle is installed in a folder “c:\kettle\kettle-latest-version”, use the following relative address “../data/mydatabase.db” or the absolute form “/kettle/data/mydatabase.db” (note the use of unix style slash).
Bug now fixed.