RE: When is %EOF not an %EOF

["Jon Paris" <Jon.Paris@xxxxxxxxxxxxxx>]

 >> Which suggests most of the performance difference may be mitigated by
using OVRDBF to enable blocking again.  I've not run any tests.

Using a OVR will have zero effect on what the RPG program does - it will
still do a single record fetch - the only difference is in the amount of
data that the OS buffers.  So, it will help some with performance but there
is a better option.

You can achieve what you want by using Block(*Yes) on the file's F-spec.
This should be used with caution as it can adversely affect the performance
but if you _know_ that you will be doing a positioning operation (CHAIN or
SETLL) followed by a series of sequential reads (e.g. filling a subfile) it
will work well for you.  Few people know about this option as it was
developed for VARPG and then added to RPG IV with very little fanfare.

Jon Paris
Partner400
www.Partner400.com