Account Options

  1. Sign in
The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Google Groups Home
« Groups Home
comp.sys.tandem
+ new post
About this group
Subscribe to this group
This is a Usenet group - learn more
Message from discussion Discrepency between SQL/MP Explain plan, actual sequence of IOs
The group you are posting to is a Usenet group. Messages posted to this group will make your email address visible to anyone on the Internet.
Your reply message has not been sent.
Your post was successful
 
From:
To:
Cc:
Followup To:
Add Cc | Add Followup-to | Edit Subject
Subject:
Validation:
For verification purposes please type the characters you see in the picture below or the numbers you hear by clicking the accessibility icon. Listen and type the numbers you hear
 
Brian Ray  
View profile  
 More options Nov 21 2007, 2:42 pm
Newsgroups: comp.sys.tandem
From: "Brian Ray" <be_...@sbcglobal.net>
Date: Wed, 21 Nov 2007 19:42:16 GMT
Local: Wed, Nov 21 2007 2:42 pm
Subject: Re: Discrepency between SQL/MP Explain plan, actual sequence of IOs
Print | View thread | Show original | Report this message | Find messages by this author

"Keith" <kd...@acm.org> wrote in message

news:2b40a60e-1477-45f1-8b07-d11188d81a4a@w34g2000hsg.googlegroups.com...

> This is pure speculation, but one thing that occurs to me is that the
> results from SQLSADISPLAY might look odd if the query was able to do
> the scan of table1 using only columns from the index.  In such a case,
> I don't know what it would report in the SQLSA.

> I'd kind of expect it to give the index name as the table name and
> tell you what it accessed from the index, but maybe it doesn't do
> that.

> I believe the Explain makes it clear whether the query plan is to do
> an index-only scan or use the index then access its base table.  Which
> does it show for your query?

> Again, this is just speculation on my part -- I don't have any
> experience with what is reported in SQLSA when the query involves an
> index-only scan.

Aha, that could be it, as the TABLE1 scan is by indexed columns only--no
base table predicate. The cursor does select un-indexed TABLE1 columns, but
only filters by indexed columns. On cursor open it could just be dealing
with TABLE1's index, which contains everything it needs for the WHERE clause
and for the inner join. Then on the actual fetch maybe it's going back and
picking up the other selected columns.

Thanks,

Brian Ray


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy
©2013 Google