Comments on: Creating a summary table in SQL Server

By: mrdenny

mrdenny — Thu, 19 Feb 2009 09:08:53 +0000

Elsuket,
Execution plans are cached when using EXEC. Depending on they query they can also be parametrized.

By: marlonr

marlonr — Wed, 18 Feb 2009 16:48:14 +0000

I think the main question here is whether he can make summary data without having to resort to cursor or dynamic SQL. There are case against and for both Cursor and Dynamic SQL. You can do a quick research on that and you’ll come up with a number of great articles. But for summarizing data, have you looked at PIVOT (see my sample here:

http://dbalink.wordpress.com/2008/10/29/summarize-aggregated-data-using-pivot-in-tsql/

) or maybe Common Table Expressions (SQL Server 2005+)?

By: elsuket

elsuket — Tue, 17 Feb 2009 22:43:56 +0000

The best way to execute dynamic T-SQL code is to execute it with through the system stored procedure sp_executesql, which is able to cache the dynamic SQL query plans (what does not do the EXEC command).

It also offers the advantage to use OUTPUT variables, and to affect them to the local variables context, as when you execute dynamic T-SQL code, its context of execution is distinct from the procedure which encapsulates it.

@++

By: mrdenny

mrdenny — Tue, 03 Feb 2009 21:40:44 +0000

The amount of data you are working with has no impact on the performance of dynamic SQL.

Dynamic SQL hurts SQL 2000 because of the way that SQL Server 2000 and older don’t handle dynamic SQL as effectively (with regard to the execution plans) as SQL 2005 and newer.

These days the big problem with dynamic SQL is that the security issues still exist (and probably always will).

By: sivu

sivu — Tue, 03 Feb 2009 05:29:11 +0000

B’cause Dynamic SQL kills the performance for large data.