Re: Improving SQL Update perfomance

Vern,

Thanks for your answer.

Nop. These are all (IIRC) tables without indexes.

I already knew the tip of removing and adding members to
the LFs but, same as you, I'm not sure if this is doable
with SQL indexes. If I have the time I'll make a couple of
tests tomorrow and let you know.

Regards,

Luis Rodriguez


--- midrange-l-request@xxxxxxxxxxxx wrote:
----------------------------------------------------------------------

message: 1
date: Wed, 28 Nov 2007 23:18:46 +0000
from: vhamberg@xxxxxxxxxxx
subject: Re: Improving SQL Update perfomance

Are there any logical files or SQL indexes based on these
fields? You could set MAINT(*DLY) on all of them and on
the physical file before doing the update, then change
back to MAINT(*IMMED) - check what the values are first.

An even surer way to eliminate index updates is to remove
all members of the LFs and delete the indexes and add
the members back afterwards and recreate the indexes -
not sure if you can remove members from indexes.

HTH
Vern

-------------- Original message ----------------------
From: Luis Rodriguez <luis.rodriguez2@xxxxxxxxx>

Hi list,

What would be the best way to improve the performance

of

a SQL Update statement executed with RUNSQLSTM?

I need to run the following SQL for several tables:

UPDATE LIBNAME/TABLE SET
Field1 = round(field1 / 1000, 2),
Field2 = round(field2 / 1000, 2),
...
Fieldn = round(fieldn / 1000, 2);

Each one of these tables can have several thousands of
records. One of them has 15 million of records, and the
process is VERY slow.

Thanks in advance,

Luis

________________________________________________________________________________

Luis Rodriguez

IBM Certified Systems Expert
eServer i5 iSeries Technical Solutions


____________________________________________________________________________________
Be a better pen pal.
Text or chat with friends inside Yahoo! Mail. See how. http://overview.mail.yahoo.com/