The size of a table. (space_used against calculated space).

Question

Post reply

The size of a table. (space_used against calculated space).

ben.brugman

SSChampion

Points: 13369
More actions
June 10, 2013 at 6:42 am

#394938

With sp_space_used you get the 'gross' size of a table. (Used_size for data)
Summing all fields with a datalength function also gives a size for a table, offcourse this is without the overhead.
The used_size is about 2 to 3 times the calculated size.
So how can I calculate the size more accurately?
Thanks for your time and attention.
ben brugman
Extra info:
This was calculated for heaps and clusters (90 percent fill rate), both were similar.
No deleted rows.
From 30 to 75 columns for each table, with a lot of NULL's.
Overhead was 3 to 5 bytes for each field.
(180 to 280 bytes of overhead for each row).

Viewing 9 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic. Login to reply

PaulB-TheOneAndOnly SSC Guru Points: 66772 More actions · Answer 1

ben.brugman (6/10/2013)
So how can I calculate the size more accurately?

The actual space used by a table is what sp_spaceused reports.

_____________________________________
Pablo (Paul) Berzukov

Author of Understanding Database Administration available at Amazon and other bookstores.

Disclaimer: Advice is provided to the best of my knowledge but no implicit or explicit warranties are provided. Since the advisor explicitly encourages testing any and all suggestions on a test non-production environment advisor should not held liable or responsible for any actions taken based on the given advice.

ben.brugman SSChampion Points: 13369 More actions · Answer 2

PaulB-TheOneAndOnly (6/10/2013)
ben.brugman (6/10/2013)
So how can I calculate the size more accurately?
The actual space used by a table is what sp_spaceused reports.

Thanks for your reply, but this does not anwser my question.

The links:

http://msdn.microsoft.com/en-us/library/ms189124.aspx

http://msdn.microsoft.com/en-us/library/ms178085.aspx

Do help, but do not even come close to the reported use using sp_spaceused.

Thanks,

ben

Dird SSCertifiable Points: 5081 More actions · Answer 3

Do you mean calculate more accurately the size before actually creating the table?

If so, at least some of inaccuracy may be that you would need to factor in the block size. If your row uses 4.5kb and you expect to have 1,000,000 rows you might have calculated 4.5gb but then because you can only fit one row within an 8k block (4.5 * 2 = too much) you end up using 1,000,000*8k blocks (about 8gb).

For smaller rows I'm not sure why it would be so far out...maybe incorrect calculations?

Indexes is easier to guess (fragmentation)~

Dird

Ron McCullough SSC Guru Points: 63877 More actions · Answer 4

ben.brugman (6/10/2013)
With sp_space_used you get the 'gross' size of a table. (Used_size for data)
Summing all fields with a datalength function also gives a size for a table, offcourse this is without the overhead.
The used_size is about 2 to 3 times the calculated size.
So how can I calculate the size more accurately?
Thanks for your time and attention.
ben brugman

This discussion of the subject may be of help to you.

http://msdn.microsoft.com/en-us/library/ms189124.aspx

If everything seems to be going well, you have obviously overlooked something.

Ron

Please help us, help you -before posting a question please read[/url]
Before posting a performance problem please read[/url]

ben.brugman SSChampion Points: 13369 More actions · Answer 5

Dird (6/10/2013)
Do you mean calculate more accurately the size before actually creating the table?

At the moment I am trying to do the calculation on existing tables. This allows me to check the method. This brings understanding and will help to calculate the size of future tables.

Ben

ben.brugman SSChampion Points: 13369 More actions · Answer 6

bitbucket-25253 (6/10/2013)
This discussion of the subject may be of help to you.
http://msdn.microsoft.com/en-us/library/ms189124.aspx

As you can see in the third message in this thread I allready mentioned exactly that link. The referenced page is actually only a very rough guide.

If everything seems to be going well, you have obviously overlooked something.

For me it's not obvious what I am overlooking, hence the question.

Because the large number of NULL's, the calculation on the web page is not very accurate. I'll try to come up with a more accurate calculation. Have to count the actual number of Nulls in the variable record fields, I'll need access to the table for this and at the moment I do not have access.

In that way I can see the amount of overhead (quite a lot) which I am missing.

Ben

Sean Pearce SSCoach Points: 15999 More actions · Answer 7

ben.brugman (6/10/2013)
If everything seems to be going well, you have obviously overlooked something.
For me it's not obvious what I am overlooking, hence the question.

Don't mistake the cheeky signature line as a rude comment.

The SQL Guy @ blogspot[/url]

@SeanPearceSQL

About Me[/url]

ben.brugman SSChampion Points: 13369 More actions · Answer 8

Sean Pearce (6/11/2013)
ben.brugman (6/10/2013)
If everything seems to be going well, you have obviously overlooked something.
For me it's not obvious what I am overlooking, hence the question.
Don't mistake the cheeky signature line as a rude comment.

Well I sure missed that one.

(English being my third language, the line was above the name and being technical also means that I tend to take things more literal than 'right' brained people.). I do like the quote though.

thanks,

ben