RE: Database monitor (V7R2) 5005 entries: meaning and QQ1000L

Hi Mandy,

Even if you don't have any thing to put in the filters pane of the
Statements window, if you have the 5005 record (or any dbmon records for
the query you want) you should be able to find the query in the unfiltered
list statement list in ACS. For example, the QQJFLD (may have to add this
via the Columns... button) is also in every dbmon record and uniquely
identifies that query in the file.

To differentiate CQE vs. SQE, use the QQSMINT5 column of the 1000 record.
https://www.ibm.com/docs/en/i/7.4?topic=view-1000-sql-information

Thank you,

Tim Clark
Dept 45X - DB2 for IBM i / SQL Optimizer

"MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxxxxxxxx> wrote on 05/09/2021
12:28:27 PM:

From: <mandy.shaw@xxxxxxxxxxxxxxxx>
To: <midrange-l@xxxxxxxxxxxxxxxxxx>,
Date: 05/09/2021 12:28 PM
Subject: [EXTERNAL] Re: Database monitor (V7R2) 5005 entries:
meaning and QQ1000L
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxxxxxxxx>

Many thanks for your input Vern and Timothy.

Timothy, I can't use the nice ACS screens, unfortunately, because the

name

of the Query being run is not logged anywhere in the Database Monitor,

so I

am having to use various bits of locally meaningful SQL to match monitor
output against specific Queries. I don't suppose there is any way of

easily

identifying the SQL for a specific panel (like you can in the

Performance

navigation screens in iNav)? That would be really useful.
(I have given up on the 5005 entries for now!)

Anyway I've now got a bit further.
Am I right Vern (I think this was your point?) that Query Sort

processing

has no specific information logged in the DBMON output, and is

effectively

entirely unpredictable in its results?
This is absolutely what I am finding - everything else seems

predictable.

In particular, if there are no sort fields at all and the optimisation
matches across SQE and CQE, then the output seems to match every time.

The

only exception to this relates to the handling of Report Breaks where

there

are no sort fields (it is clear that SQL GROUP type processing is not

being

used, either in SQE or in CQE).

Finally, another question (sorry): what is the 'proper' way of

identifying

SQE vs CQE in the DBMON output? I am using

QQSMINTF/Plan_Iteration_Number,

treating null as CQE and everything else as SQE - this seems to work,

but am

I missing a more formal/accurate method? Not all the record types

include

the query options library name, or I'd use that.

Many thanks
Mandy