Re: SQL versus READ for sequential processing

On Tue, 2008-10-14 at 17:14 -0500, Vern Hamberg wrote:

Was hoping someone would keep this going!
Yes - RTFM showed me that DYNSLT enables access path sharing if the
access path being shared has no S/O rules - otherwise, the S/O rules
have to be the same, if not using DYNSLT.

Now I think David really does need both - he needs the S/O for his RPG
stuff, and he wants to use SQE for his SELECT statement. Making the
non-S/O logical (or index) just lets him take advantage of access path
sharing, if he so chooses. Just has to change a QAQQINI file.

It's always good to save space with access path sharing - also takes
less time to maintain indexes.

As I know you know!

Regards
Vern

I was tempted to hold forth on the performance and space implications of
DYNSLT, but feared to tread toward the root-of-all-evil <grin />. I
shall not even mention the MAINT() parameter of a LF. (Can you say
"paralipsis", boys and girls?)

Cheers,
Terry.

[previous discussion snipped]