Hi Irwan,

I had to do a :

FileCurrentSize = (cast(SF.size as bigint) * 8)/1024,

in order to avoid an integer overflow.

Best regards,

Henrik Staun Poulsen

http://www.stovi.com

hi Jeff and Peter,

Thank you very much for a super solution.

Best regards,

Henrik

Hi Ron,

I have seen this situation too.

But I've also found that even if the USER_Scans column was zero, the SYSTEM_scans was not zero.

I'm not sure why the System_Scans column exists....

Hi Kingston,

I would check my code for the occurancy of any "/".

When I found one, I would change the code like this:

from:

Nominator / Denominator

to:

Nominator /...

Hi Rajat,

According to dbtuna you are just waiting for the harddisk to repond. Other than getting a faster hard disk (SSD?), I have no suggestions.

I have just been thru a...

Hi Rajat,

My guess is that your table did a lot of page splits, resulting in a lot of unused space. When you rebuild your clustered index, the table size drops...

I've checked the link to SQLServerPedia, but I still disagree that a page in a heap is splitted, just because a row is longer.

I wrote this example:

BEGIN TRY

DROP TABLE tmp.hsp...

3 million rows is nothing. I have a table with 60 billion row of a similar design.

But this table is not used like your usage.

I have 6 other tables...

It is a lot easier to keep a customer, than to get a new one.

Similar thing must be true for contributors.

If every single person on the list just wrote one...

The code does not take the instance name into account.

Here is a solution that does.

/******************************************************************************

http://www.sqlservercentral.com/scripts/Performance+Monitoring/71677/

NOTES:

This script provides a sampling of PLE based on (1) minute...

I particular like the "DO NOTE" comment.

I got burnt by that a few years ago.

Hej Lars,

For SQL Server 2008 you need to change line 1

from

SELECT crdate FROM sysdatabases WHERE name = 'tempdb'

to

SELECT create_date FROM sys.databases WHERE name = 'tempdb'

Med venlig hilsen

Henrik Staun...

I think that sp_indexinfo is much better than this query.

'31-Jan-2010'

It that the correct (international) way to hard-code dates?

I think I would have used '2010-01-31' or is it '20100131' to ensure that it will work around the globe?

Hugo,

<<If you want to index a computed column, you must persist it.

Are you sure?

I've just added a computed column to one of our big tables (+4 billion rows).

I did not...