[dba-SQLServer] Does building cover indexes pay?

Mark Breen marklbreen at gmail.com
Sat Jun 12 03:24:50 CDT 2010


Hello John,

I agree with Michael here, only one persons opinion matters here and that's
yours as only you really now the true effect.

Is it not possible to do a method A and Method B comparison?

I am presuming that each time you face the decision the results are
significantly different that it is difficult to identify the true answer.
 If that is the case, the again, only you will really know.

Can I ask, have you read up much on ETL?  The works you are doing nowadays
is AFAICS, ETL and I believe that there are plenty of ETL tools.  In fact it
is a subject area in itself and my instinct tells me that traditional
approaches to data management is different when you are doing ETL.

I only saw it for a few months when I worked for a while in an Oracle Shop,
but they had a whole ETL department with their own tools.  Those tools did a
lot of the heavy lifting that you are coding.

HTH somewhat

Mark


On 10 June 2010 01:57, Michael Maddison <michael at ddisolutions.com.au> wrote:

> Hi John,
>
> I think you will only get opinions, if that.
> You are going to have to test each scenario IMO.
> Another option for testing is to enable/disable the covering indexes.
> You probably need to update statistic after enabling if the data has
> changed.
> If the data hasn't changed then this 'should' be your best option.
>
> Cheers
>
> Michael M
>
>
> I have processes which process huge tables (65 million records are
> common) and update specific
> fields with the results of user defined functions.  For example I have a
> function which builds a
> text "code" using four passed in values.  These four values come from
> four fields of the record and
> the resulting code is placed back in another field of the same record.
>
> ATM I build a "cover index" just before updating this field, perform the
> update, then delete the
> index when done.
>
> I am wondering whether you guys believe that the time to build the cover
> index will be recovered by
> less time taken to perform this update.  IOW does it take sufficiently
> less time to get data from an
> index vs a table scan that it is worth building the index?
>
> --
> John W. Colby
> www.ColbyConsulting.com
> _______________________________________________
> dba-SQLServer mailing list
> dba-SQLServer at databaseadvisors.com
> http://databaseadvisors.com/mailman/listinfo/dba-sqlserver
> http://www.databaseadvisors.com
>
>
> No virus found in this incoming message.
> Checked by AVG - www.avg.com
> Version: 9.0.829 / Virus Database: 271.1.1/2926 - Release Date: 06/10/10
> 04:35:00
>
> _______________________________________________
> dba-SQLServer mailing list
> dba-SQLServer at databaseadvisors.com
> http://databaseadvisors.com/mailman/listinfo/dba-sqlserver
> http://www.databaseadvisors.com
>
>



More information about the dba-SQLServer mailing list