(Environment: Adaptive Server 11.5 on Windows NT)
I have got a scenario where, I have to read of whole record from an existing
table(60 columns wide) and write it back. This are random reads from the
I have tried DBLib updatable cursors and Stored procedures, but neither
gives me satisfactory performance.
Has anyone got better idea to deal with the case.
Thanks in advance.
Subject: Best solution for the problem.
Date: Mon, 9 Mar 1998 15:55:10 -0000
X-Newsreader: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Xref: forums-1-dub sybase.public.sqlserver.nt:5125
Article PK: 1092489
Date: Mon, 09 Mar 1998 16:44:02 -0700
From: Bret Halford <firstname.lastname@example.org>
Organization: Customer Service & Support
X-Mailer: Mozilla 3.0 (X11; U; SunOS 5.5.1 sun4m)
To: Sarju Patel <email@example.com>
Subject: Re: Best solution for the problem.
Content-Type: text/plain; charset=us-ascii
Xref: forums-1-dub sybase.public.sqlserver.nt:5124
Article PK: 1092486
Could you be more specific about your results?
What performance are you getting?
What performance do you need?
Try it first as straight sql,
ie select * from mytable where <primary_key> = <value>.
What kind of performance are you getting there?
If it is not adequate, does showplan show proper index use,
or is a table scan occurring? Could you improve performance
with a better index?
What performance do you get when inserting/updating the row?
Again, what does showplan show? Could you improve
performance by reducing the number of indexes on the table?
Do the columns chosen for the clustered index spread writes
out, or does it cause a point or points of contention?
Bret Halford Imagine my disappointment
Sybase Technical Support in learning the true nature
3665 Discovery Drive of rec.humor.oracle...
Boulder, CO 80303