Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
 
Log in  ::  Register  ::  Not logged in
 
 
 
        
Home       Members    Calendar    Who's On


Add to briefcase ««123»»

Finding rows count in a table without Select... Expand / Collapse
Author
Message
Posted Friday, March 26, 2010 10:05 AM
SSC Journeyman

SSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC Journeyman

Group: General Forum Members
Last Login: Monday, January 21, 2013 2:17 PM
Points: 98, Visits: 83
From my testing of these scripts, the row counts are only accurate if the table has a clustered index.

Gary
Post #890791
Posted Friday, March 26, 2010 10:16 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Sunday, December 14, 2014 5:56 PM
Points: 114, Visits: 459
I believe the WHERE clause here:
Where indid < 2
means that it's on the clustered index. So wouldn't it alway be accurate regardless of statics, since that's the actual physical layout of the rows?

Let me know if that's not correct.

Thanks.




G. Milner
Post #890809
Posted Friday, March 26, 2010 10:21 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Sunday, December 14, 2014 5:56 PM
Points: 114, Visits: 459
Hi, Manish.
I turned an email I wrote -- based on your post -- to a colleague (ETL designer) into a blog post on my site here, and cited you. Hope that's OK.

Thanks.




G. Milner
Post #890818
Posted Friday, March 26, 2010 10:28 AM
SSC Journeyman

SSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC JourneymanSSC Journeyman

Group: General Forum Members
Last Login: Monday, January 21, 2013 2:17 PM
Points: 98, Visits: 83
Regarding the clause "Where indid < 2", if you have a clustered index there will be a row in sysindexes with indid = 1. If you do not have a clustered index, there will be a row in sysindexes = 0.

From my testing, only the tables with a clustered index (indid = 1) are accurate.

Gary
Post #890832
Posted Friday, March 26, 2010 9:16 PM


SSC-Dedicated

SSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-Dedicated

Group: General Forum Members
Last Login: Today @ 2:14 PM
Points: 35,770, Visits: 32,435
If you need more accurate results, try using DBCC UPDATEUSAGE first. If you haven't done it in awhile, it could take some good bit of time and will take less time the more you use it.

--Jeff Moden
"RBAR is pronounced "ree-bar" and is a "Modenism" for "Row-By-Agonizing-Row".

First step towards the paradigm shift of writing Set Based code:
Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column."

(play on words) "Just because you CAN do something in T-SQL, doesn't mean you SHOULDN'T." --22 Aug 2013

Helpful Links:
How to post code problems
How to post performance problems
Post #891224
Posted Saturday, March 27, 2010 1:24 AM
SSC Veteran

SSC VeteranSSC VeteranSSC VeteranSSC VeteranSSC VeteranSSC VeteranSSC VeteranSSC Veteran

Group: General Forum Members
Last Login: 2 days ago @ 4:44 AM
Points: 290, Visits: 1,082
DBCC UPDATEUSAGE - doesn't that only have to be run on SQL 2000 or upon upgrade from 2000 -> 2005?

I thought the issue had been corrected in 2005 onwards?

DBCC UPDATEUSAGE BOL quote:

"In SQL Server 2005, these values are always maintained correctly. Databases created on SQL Server 2005 should never experience incorrect counts, however, databases upgraded to SQL Server 2005 may contain invalid counts."

As I mentioned before, using sys.dm_db_partition_stats is the most reliable, future-proof method.

Chris
Post #891263
Posted Saturday, March 27, 2010 12:58 PM


SSC-Dedicated

SSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-Dedicated

Group: General Forum Members
Last Login: Today @ 2:14 PM
Points: 35,770, Visits: 32,435
Chris Howarth-536003 (3/27/2010)
DBCC UPDATEUSAGE - doesn't that only have to be run on SQL 2000 or upon upgrade from 2000 -> 2005?

I thought the issue had been corrected in 2005 onwards?

DBCC UPDATEUSAGE BOL quote:

"In SQL Server 2005, these values are always maintained correctly. Databases created on SQL Server 2005 should never experience incorrect counts, however, databases upgraded to SQL Server 2005 may contain invalid counts."

As I mentioned before, using sys.dm_db_partition_stats is the most reliable, future-proof method.

Chris


Actually, that's exactly correct, Chris. I've been stuck in a 2k world for much too long. Being in a bit of a hurry, all I saw was a reference to sysindexes and the word "inaccurate" and didn't realize folks were talking of 2k5+.


--Jeff Moden
"RBAR is pronounced "ree-bar" and is a "Modenism" for "Row-By-Agonizing-Row".

First step towards the paradigm shift of writing Set Based code:
Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column."

(play on words) "Just because you CAN do something in T-SQL, doesn't mean you SHOULDN'T." --22 Aug 2013

Helpful Links:
How to post code problems
How to post performance problems
Post #891326
Posted Monday, March 29, 2010 2:11 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Tuesday, December 7, 2010 2:34 AM
Points: 13, Visits: 31
DBCC UPDATEUSAGE - doesn't that only have to be run on SQL 2000 or upon upgrade from 2000 -> 2005?


Actually that may provide the explanation. My original post was about inaccurate counts in a 2000 database I had migrated to 2008 and I didn't know about the caveat, so that may well be the answer.

Thanks Jeff, Chris (and All)



Post #891641
Posted Wednesday, March 31, 2010 12:42 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: 2 days ago @ 3:48 AM
Points: 129, Visits: 199
Hello
This is the cool example, but you do not need to write own procedure to determine the tables row count. Just try this:

exec SP_MSFOREACHTABLE @command1=N' exec SP_SPACEUSED ''?'' '

Regards
Yankov Yanko
Post #893502
Posted Friday, April 9, 2010 12:35 PM
Forum Newbie

Forum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum Newbie

Group: General Forum Members
Last Login: Friday, April 30, 2010 11:50 AM
Points: 1, Visits: 24
And th equestion again: How recent are your statistics, indexes, etc. ? If you can not tell, better use count(*) to get the numbers. It is slow but accurate.
Post #900766
« Prev Topic | Next Topic »

Add to briefcase ««123»»

Permissions Expand / Collapse