By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
440,640 Members | 1,598 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 440,640 IT Pros & Developers. It's quick & easy.

Temp Table Faster?

P: n/a
If you were doing paging of results on a web page and were interested
in grabbing say records 10-20 of a result set. But also wanted to know
the total # of records in the result set (so you could know the total #
of pages in the set).

Would it be better to query the DB table 2X. Once for Count(*). And
again for the records for the current page?

Or better to create a temp table, select the records into it, and then
get count(*) and the page results from the temp table?

I saw an example in a book that made a temp table to do this and to me
it seemed like it would be slower. I don't get the reason for a temp
table. Anyone have any ideas?

Dec 23 '05 #1
Share this Question
Share on Google+
5 Replies


P: n/a
wa********@yahoo.com wrote:
If you were doing paging of results on a web page and were interested
in grabbing say records 10-20 of a result set. But also wanted to know
the total # of records in the result set (so you could know the total #
of pages in the set).

Would it be better to query the DB table 2X. Once for Count(*). And
again for the records for the current page?

Or better to create a temp table, select the records into it, and then
get count(*) and the page results from the temp table?

I saw an example in a book that made a temp table to do this and to me
it seemed like it would be slower. I don't get the reason for a temp
table. Anyone have any ideas?


Take a look here:
http://www.aspfaq.com/show.asp?id=2120

--
David Portas
SQL Server MVP
--

Dec 23 '05 #2

P: n/a
Yeah, I see in 2000 you'd do an insert into a temp table to assign a #
to each row. In 2005 this is not necessary, so would there be any
reason speed wise to use a temp table?

Dec 23 '05 #3

P: n/a
(wa********@yahoo.com) writes:
If you were doing paging of results on a web page and were interested
in grabbing say records 10-20 of a result set. But also wanted to know
the total # of records in the result set (so you could know the total #
of pages in the set).

Would it be better to query the DB table 2X. Once for Count(*). And
again for the records for the current page?

Or better to create a temp table, select the records into it, and then
get count(*) and the page results from the temp table?

I saw an example in a book that made a temp table to do this and to me
it seemed like it would be slower. I don't get the reason for a temp
table. Anyone have any ideas?


A temp table could be slower because of recompilations.

An alternative is to use a permanent table, that would have some session
key and an IDENTITY column (in SQL 2000). When the user makes his first
search, you get all data into that table. Then as he pages on, you retrieve
the rows from this table. This means you don't have to redo the query for
subsequent pages, but can get it from the table. This is likely to give
better performance, and another advantage: a fixed result. If the result
can change as the user browse, he may miss a row that initially was row
101, but now is row 100.

Finally, don't design pages where the user only can get 10 rows at a
time. I hate those. Give me at least 100 at a time.

--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/pro...ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinf...ons/books.mspx
Dec 23 '05 #4

P: n/a

Erland Sommarskog wrote:

An alternative is to use a permanent table, that would have some session
key and an IDENTITY column (in SQL 2000). When the user makes his first
search, you get all data into that table. Then as he pages on, you retrieve
the rows from this table. This means you don't have to redo the query for
subsequent pages, but can get it from the table. This is likely to give


Depending on the actual practical needs, yet another variation of this
technique is to save only the primary keys into the permanent table.
Requires less disk space and displays changes made after the PK set was
materialized.

Dec 25 '05 #5

P: n/a
Alexander Kuznetsov (AK************@hotmail.COM) writes:
Erland Sommarskog wrote:

An alternative is to use a permanent table, that would have some
session key and an IDENTITY column (in SQL 2000). When the user makes
his first search, you get all data into that table. Then as he pages
on, you retrieve the rows from this table. This means you don't have to
redo the query for subsequent pages, but can get it from the table.
This is likely to give


Depending on the actual practical needs, yet another variation of this
technique is to save only the primary keys into the permanent table.
Requires less disk space and displays changes made after the PK set was
materialized.


Good point. There is a potential problem, though, if rows can be deleted.
(But this could be indicated when returning the data.)

There is also a risk for confusion, if the user selects data to be
sorted by something which is not in the key, for instance price, and the
price is updated while the user is paging.

What this really boils down to is that to implement paging properly, you
need to understand the business domain.
--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/pro...ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinf...ons/books.mspx
Dec 25 '05 #6

This discussion thread is closed

Replies have been disabled for this discussion.