[KLUG Programming] SQL question

[Adam Tauno Williams]

> > The most important databases (of course) are huge though, both in the
> > number of fields and the number of records (well huge comparatively).
> >  One of the databases has >1400 fields in it.  About 400 fields are
> Yikes!
> > for that!).  Another one is like 167 fields.  Access chokes on them
> > with ODBC and an export just is not reliable.  The FileMaker ODBC
> Way too big for toy tools.  Can you get JDBC access to the FileMaker data?

If you have the Sun JVM installed there is a JDBC/ODBC bridge driver
that lets JDBC apps connect to any ODBC datasource.  This works better
with some ODBC drivers that others, but is work a try.  The class name
is "sun.jdbc.odbc.JdbcOdbcDriver", and the database URL looks like
"jdbc:odbc:{DSN Name}".

I'd test connectivity with a generic JDBC app like DbVisualizer
(http://www.dbvis.com/products/dbvis/), and if that seems reliable, I'd
then try opendbcopy.  Mostly because the error messages generated by
opendbcopy when something doesn't work aren't usually terribly helpful.

> If so use - http://opendbcopy.sourceforge.net/index.html
> Designed for unloading/loading/migrating real databases.

Since you've got lots of data it is probably recommended that you crank
up the amount of memory the JVM will let itself allocate (-Xmx512m for
example to let it use 512Mb of RAM).