RE: CQE, SQE and QAQQINI

This is the only known detractor I have seen to making the switch. CQE can
use keyed S/O LFs and since in BPCS keys queries were written with these S/O
LFs in mind, they're tightly integrated and perform well. SQE cannot use S/O
LFs and even if an equivalent full sql index exists, it won't perform as
well (it'll come close). If you don't build full sql indexes (or non-S/O
keyed LFS) though, CQE will beat the pants off of SQE, since SQE will have
to default to table scans in that case.

That said, IGNORE_DERIVED_INDEX is a default on V6R1 and I have seen many
fantastic results on customers machines. One guy even got a promotion after
he made the switch (no kidding; system performance went up dramatically and
CPU usage dropped dramatically).

Elvis

Celebrating 11-Years of SQL Performance Excellence on IBM i, i5/OS and
OS/400
www.centerfieldtechnology.com


-----Original Message-----
Subject: Re: CQE, SQE and QAQQINI

Would it concern you that a major software house sides with your boss?
http://www-912.ibm.com/n_dir/nas4apar.nsf/c79815e083182fec862564c00079d117/2
ed784e875e77a4586257490003c6d39?OpenDocument

********************************************************
SQL Performance
********************************************************
--------------------------------------------------------
Affected BPCS versions: V6.x, V8.x
--------------------------------------------------------
Infor recommends the following QAQQINI file setting to
be activated for support with R610:
QAQQINI IGNORE_DERIVED_INDEX *NO

The BPCS and ERPLX database contains large numbers of
Select/Omit logical files and if this option is not set,
these logical files will be ignored when the SQL Optimzer
selects logical files to run a given query. This can cause
poor performance.

The following command shows how to setup this option...


Rob Berendt