August 18, 2013 at 9:47 pm
Hi All,
I was trying to use partitioned index for a slow query but I was stopped with following response from my manager :
"There are few cases where partitioning seems logical and looks promising but then we found that with right set of indexes and tweaks, performance was equal or even better as any partition scheme would otherwise offer.
I think we had external SQL consultants who were of same opinion we well."
If you please share your experience with partitioning, it would be a great help.
Regards
August 19, 2013 at 7:30 am
This depends on the query and your data. Can you provide more details?
August 19, 2013 at 7:45 am
T.Ashish (8/18/2013)
Hi All,I was trying to use partitioned index for a slow query but I was stopped with following response from my manager :
"There are few cases where partitioning seems logical and looks promising but then we found that with right set of indexes and tweaks, performance was equal or even better as any partition scheme would otherwise offer.
I think we had external SQL consultants who were of same opinion we well."
If you please share your experience with partitioning, it would be a great help.
Regards
Your boss is absolutely correct. You should NOT be looking to partitioning to primarily speed up a slow query. That is NOT what it was built for. It CAN provide incredible performance gains - but you can also REALLY mess things up and cause pain for yourself, especially if you are not experienced with partitioning in the first place and don't get professional help with it.
TUNE FIRST (and there are a TON of things to be evaluated/refactored/done when tuning), and only THEN should you look to alternatives to get additional performance gains if needed.
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
August 19, 2013 at 8:02 am
https://www.simple-talk.com/sql/database-administration/gail-shaws-sql-server-howlers/
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
August 19, 2013 at 8:51 am
Kevin, Gail
Thanks a lot.
August 19, 2013 at 9:25 pm
One more Issue:
Is it going to change in next version?
August 19, 2013 at 11:50 pm
It really depends on the access pattern of your data through queries. For generic queries, it is always better to check performance improvement with indexes.
IMO Go for partition if
1. The query access pattern always needs a where condition on partition columns. Remember post partition, to get better performance the query always should use partition column else it might result in table scan
2. The data you are handling is very high. The partition helps in purging / archival of data from certain partitions using ALTER SWITCH PARTITION (DBA perspective)
August 20, 2013 at 3:43 am
T.Ashish (8/19/2013)
One more Issue:Is it going to change in next version?
Partitioning is not primarily for performance. It's for management, data loads, easy archiving. That it doesn't magically improve performance is not a bug to be fixed.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
August 20, 2013 at 6:34 am
Gail,
I understood that. And I have read the article you mentioned, its really nice.
Actually I meant this feature of SQL Server 2014
a. Online Indexing at partition level
b. Statistics at the partition level
I found it on some blog.
August 20, 2013 at 7:07 am
T.Ashish (8/20/2013)
Gail,I understood that. And I have read the article you mentioned, its really nice.
Actually I meant this feature of SQL Server 2014
a. Online Indexing at partition level
b. Statistics at the partition level
I found it on some blog.
I've not personally checked but Paul White (whom I have quite a bit of trust in), says that online indexing is in 2014 CTP1. Unfortunately, I've not heard of a retrofit to any other versions.
--Jeff Moden
Change is inevitable... Change for the better is not.
Viewing 10 posts - 1 through 9 (of 9 total)
You must be logged in to reply to this topic. Login to reply