Re: Giant proxy DBD for multiple databases

[dws]

It seems to me that this should be possible with SQL::Statement or DBD::AnyData or something similar, and a lot of work creating DBD::AnyData drivers for each of the back-end types (VSAM, flatfile dataset, M204, PDS, ...) that submit a simplified version of the query to the COBOL driver routines on the various hosts.

If I were unable to find something off-the-shelf, subclassing or modifying SQL::Statement the direction I'd go. It's relatively easy to parse an extended SQL grammar. It's harder to build an execution plan that handles heterogenous, distributed data. That's where I suspect you'll be spending the bulk of your time.

For reference, our "small" test dataset is about 100 billion rows in DB/2 on the 'frame.

I'm saving that line as a reminder to have people quantify what they mean by "small". :)