Re: Index Advisor Question

IBM would say that point 1 really isn't true any more under most circumstances.

Point 2 is not completely true since any subsequent query that needs the "front part" of the new key set should be able to use it.

Point 3 should already have been taken into account by the advisor since it is a time-based calculation.

I think your first comment is probably the best answer "Old habits based on old experience ..." <grin>


On 2013-09-18, at 6:20 PM, Gary Thompson <gthompson@xxxxxxxxxxx> wrote:

Evan,
Yes, just the question I had for myself and the reason I posted the way I did.

Old habits based on old experience maybe, but my thinking was/is:
1) indexes are expensive
2) multiple column indexes are relatively
less useful because fewer queries benefit
3) diminishing returns as added columns have
fewer and fewer "candidate rows"


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Evan Harris
Sent: Wednesday, September 18, 2013 4:11 PM
To: Midrange Systems Technical Discussion
Subject: Re: Index Advisor Question

I'm curious as to why those who have offered alternatives would not take the index advisor's advice.

Just looking to understand how a human expert discounts the advice of a code-driven tool and hoping to gains some insight into how to use index advisor better.

Why a 10 column index would be a red flag would be a good start - why is that specifically bad ? (Wanting to understand why that is so, not question the statement per se)


On Thu, Sep 19, 2013 at 8:54 AM, William P Hunter <wphunter@xxxxxxxxxxx>wrote:

I would start debug and run the sql statement via strsql and see what
shows up in the jobs messages.

From: dieter.bender@xxxxxxxxxxxx
To: midrange-l@xxxxxxxxxxxx
Subject: Re: Index Advisor Question
Date: Wed, 18 Sep 2013 22:45:44 +0200

Hi Gary,

first of all, I would collect some data with database Monitor
(STRDBMON)

to have a look to the real cost of those queries, then trying your
suggestion and have a look again to the cost in real life. If the cost
(= response time) is fine, best decision is: just do nothing.

Dieter
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)
mailing

list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please
take a moment to review the archives at
http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take
a moment to review the archives at
http://archive.midrange.com/midrange-l.

--

Regards
Evan Harris
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at http://archive.midrange.com/midrange-l.



--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

Jon Paris

www.partner400.com
www.SystemiDeveloper.com