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

Optimizing Cursor Performance Expand / Collapse
Author
Message
Posted Sunday, July 24, 2005 6:17 PM


SSC-Dedicated

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

Group: Administrators
Last Login: Today @ 8:31 AM
Points: 33,155, Visits: 15,286

Apologies for this. I have removed the article. If links are broken, that it why.

 








Follow me on Twitter: @way0utwest

Forum Etiquette: How to post data/code on a forum to get the best help
Post #203802
Posted Saturday, November 21, 2009 7:11 AM
Forum Newbie

Forum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum Newbie

Group: General Forum Members
Last Login: Monday, November 23, 2009 12:58 PM
Points: 2, Visits: 6
hi @all,

i want to know a bit more about cursors and the technical way they are realized. my understanding is, that cursors are pointers on a result-set within the ram. the result-set is small (comparing to the table the rows are from) and the sql which perform the result-set is executed only one time (i hope :o) ), right?
So what we have is a small result-set within a location the cpu has the fastest way to grab rows from this result-set.
why should this cursors (used in for loops) not be more efficient than ordinary sql-queries (in for loops)?

ordinary queries grap the rows from the hard-disk, where the cpu needs more time to grap them. and erery time the sql-query is executed the a full table-scan must be performed to find the row.

abstract:
cursors (used in for loops):
- sql-query to get the result-set is executed one time, to make a result set (in the initialsation-part of procedure)
- one small result-set
- fastest way for a cpu to grap the data, because they are in the ram

ordinary queries (used in for loops):
- executed to get on row per query
- in every execution the whole table has to be scaned to get one row, so there is no result-set to crawl over
- no fast way for the cpu to grap the data, because data is on harddrive

what du you think about what my understanding of cursors is. am i right?
please let me know.

cheers tommy



Post #822880
Posted Saturday, November 21, 2009 7:59 PM


SSC-Insane

SSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-InsaneSSC-Insane

Group: General Forum Members
Last Login: Today @ 10:43 AM
Points: 23,215, Visits: 31,895
You should know that wether you are using cursors or a while loop it is still RBAR (a Modenism for Row By Agonizing Row) processing. In almost all cases, converting cursor-based solutions to set-based solutions will result in improved performance and scalability of your code.



Lynn Pettis

For better assistance in answering your questions, click here
For tips to get better help with Performance Problems, click here
For Running Totals and its variations, click here or when working with partitioned tables
For more about Tally Tables, click here
For more about Cross Tabs and Pivots, click here and here
Managing Transaction Logs

SQL Musings from the Desert Fountain Valley SQL (My Mirror Blog)
Post #822949
Posted Sunday, November 22, 2009 3:43 PM


SSC-Dedicated

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

Group: General Forum Members
Last Login: Today @ 9:09 AM
Points: 36,941, Visits: 31,443
tho.pfaff (11/21/2009)
hi @all,

i want to know a bit more about cursors and the technical way they are realized. my understanding is, that cursors are pointers on a result-set within the ram. the result-set is small (comparing to the table the rows are from) and the sql which perform the result-set is executed only one time (i hope :o) ), right?
So what we have is a small result-set within a location the cpu has the fastest way to grab rows from this result-set.
why should this cursors (used in for loops) not be more efficient than ordinary sql-queries (in for loops)?

ordinary queries grap the rows from the hard-disk, where the cpu needs more time to grap them. and erery time the sql-query is executed the a full table-scan must be performed to find the row.

abstract:
cursors (used in for loops):
- sql-query to get the result-set is executed one time, to make a result set (in the initialsation-part of procedure)
- one small result-set
- fastest way for a cpu to grap the data, because they are in the ram

ordinary queries (used in for loops):
- executed to get on row per query
- in every execution the whole table has to be scaned to get one row, so there is no result-set to crawl over
- no fast way for the cpu to grap the data, because data is on harddrive

what du you think about what my understanding of cursors is. am i right?
please let me know.

cheers tommy


I'm not sure who wrote that for you but it's mostly full of hooie... not all cursors are in ram just like not all non cursors are on disk.

Baiscally, cursors will always be slower than "Set Based" queries because cursors overcome and overwhelm the "natural" way that SQL Server operates.

With only 1 exception in 10,000, you should simply avoid cursors.


--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 #823023
Posted Sunday, November 22, 2009 4:32 PM


SSCommitted

SSCommittedSSCommittedSSCommittedSSCommittedSSCommittedSSCommittedSSCommittedSSCommitted

Group: General Forum Members
Last Login: Monday, February 11, 2013 8:27 AM
Points: 1,735, Visits: 1,947
tho.pfaff (11/21/2009)
ordinary queries grap the rows from the hard-disk, where the cpu needs more time to grap them. and erery time the sql-query is executed the a full table-scan must be performed to find the row.

It is not always true that ordinary queries fetches the rows from hard disk.

Regarding full table scan must be performed to find a row, that would be true if the company has a policy of no indexes and would never use an index .... .

Your assumption is wrong Tommy, a full table scan happens when there are no indexes and can be avoided by having appropriate indexes..



Bru Medishetty

Blog -- LearnSQLWithBru

Join on Facebook Page Facebook.com\LearnSQLWithBru

Twitter -- BruMedishetty
Post #823033
Posted Monday, November 23, 2009 12:45 PM


Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Wednesday, July 23, 2014 8:16 AM
Points: 1,035, Visits: 410
Bru Medishetty (11/22/2009)
tho.pfaff (11/21/2009)
ordinary queries grap the rows from the hard-disk, where the cpu needs more time to grap them. and erery time the sql-query is executed the a full table-scan must be performed to find the row.

It is not always true that ordinary queries fetches the rows from hard disk.

Regarding full table scan must be performed to find a row, that would be true if the company has a policy of no indexes and would never use an index .... .

Your assumption is wrong Tommy, a full table scan happens when there are no indexes and can be avoided by having appropriate indexes..


What you said is true, but it seems to me that the question wasn't between a set-based query and cursor, but between a query in a for loop thus making it a RBAR operation rather than set-based.

The question seems to spring from a commonly held belief that if you need a RBAR operation (you almost never do) you are better off "rolling your own" cursors by using a loop that by using a cursor. My testing of this has proven (to me) that if you really need RBAR operation, you are likely better off using a real cursor as opposed to "rolling your own."



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

If most people are not willing to see the difficulty, this is mainly because, consciously or unconsciously, they assume that it will be they who will settle these questions for the others, and because they are convinced of their own capacity to do this. -Friedrich August von Hayek



*****************/
Post #823463
Posted Monday, November 23, 2009 5:41 PM


SSC-Dedicated

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

Group: General Forum Members
Last Login: Today @ 9:09 AM
Points: 36,941, Visits: 31,443
I agree with that whole heartedly, DC. I've found that a well written cursor will frequently beat a home grown While Loop just as you said. I'll also add that I think of While Loops as nothing more than "home grown cursors" and have resorted to calling them "cursors" even though that's probably a bit confusing to most. Based on that, allow me to correct what I said in the post previous to this one just to be clear... there are exceptions to every rule but all forms of RBAR (not just cursors) should be avoided if possible. I'll also add that, given a little thought, it's usually possible to avoid them without resorting to just another form of RBAR.

--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 #823579
« Prev Topic | Next Topic »

Add to briefcase «««123

Permissions Expand / Collapse