Re: Format Level ID

Rob,

This is the crux of the issue.

IBM has never fully documented exactly how the "hash" function in QDBLVLID that computes the record format level ID really "works."      And we all know that there have been "bugs" from time to time, over the years, with various PTFs, etc., to "fix" the problems.

The OP in this case says he was following IBM documentation where you can supposedly create a LF with exactly the same fields and in the same positions as the original file, and it is supposed to end up with the same exact record format level ID.   


For example, this was a common technique during the Y2K conversion period, and also more recently, for "modernizing" your database, where you want to replace a DDS-defined PF with an SQL DDL defined table, and then create a LF over that PF, that remains "backwards-compatible" with the original PF, so you do not have to recompile every single program that used that PF.

Just saying ...

Mark S. Waterbury






On Tuesday, November 12, 2019, 2:58:50 PM EST, Rob Berendt <rob@xxxxxxxxx> wrote:

IDK if the underlying file(s) they are built on make a difference.
For example if I take a file called IIM and copy it to IIX and compile the same LF DDS against both will both LF's have the same record id?  Now, since the OP's first file was against a PF and the second one was a Join LF might that be the reason for the different record id?
Or did no one else not wonder that?

Rob Berendt