IT SPACE: Knowledge

Cumulative update package 3 for SQL Server 2005 Service Pack 2

Wed, 22 Aug 2007 21:37:10 GMT

Microsoft has announced the cumulative update package 3 for SQL Server 2005 Service Pack 2.

You can find usefull information here: http://support.microsoft.com/kb/939537

CTE

Fri, 27 Jul 2007 12:00:06 GMT

A Common Table Expression (CTE) is similar to a view or derived query, allowing you to create a temporary query that can be referenced within the scope of a SELECT, INSERT, UPDATE, or DELETE query. Unlike a derived query, you don’t need to copy the query definition multiple times each time it is used. You can also use local variables within a CTE definition—something you can’t do in a view definition.

WITH expression_name [ ( column_name [ ,...n ] ) ]

AS ( CTE_query_definition )

A non-recursive CTE is one that is used within a query without referencing itself. It serves as a temporary result set for the query. A recursive CTE is defined similarly to a non-recursive CTE, only a recursive CTE returns hierarchical self-relating data. Using a CTE to represent recursive data can minimize the amount of code needed compared to other methods.

Examples - soon ;)

RECOMPILE option

Fri, 27 Jul 2007 00:30:07 GMT

SQL Server 2005 has introduced statement-level stored procedure recompilation. Now instead of an entire stored procedure recompiling when indexes are added or data is changed to the referenced tables, only individual statements within the procedure impacted by the change are recompiled. Typically, you will want to use this RECOMPILE query hint within a stored procedure—so that you can control which statements automatically recompile—instead of having to recompile the entire stored procedure.

SELECT SalesOrderID,ProductID,UnitPrice,OrderQty

FROM Sales.SalesOrderDetail

WHERE UnitPrice > @Price

ORDER BY SalesOrderID,ProductID

OPTION (RECOMPILE)

DDL triggers syntax

Wed, 18 Jul 2007 09:27:30 GMT

DDL triggers are designed using the following syntax:

CREATE TRIGGER trigger_name
ON { ALL SERVER | DATABASE }
[ WITH [ ENCRYPTION ] | [ EXECUTE AS CALLER | SELF | 'user_name' ]
{ FOR | AFTER } { event_type | event_group } [ ,...n ]
AS { sql_statement  [ ; ]  [ ...n ] }

The On clause controls the scope of the trigger—it is tied to database or server events. Like stored procedure triggers, DDL triggers can be encrypted and executed alternatively by caller, schema owner, or specific database user.

DDL triggers - EventData()

Mon, 16 Jul 2007 22:08:41 GMT

Since DDL triggers are not tied to table modifications, as in the case of DML triggers, they do not have Inserted and Deleted virtual tables. Instead, you can get information about the DDL modification statement that has fired the trigger using a special EventData() function.When the statement that modifies a table is issued, EventData() will return data in the form of an XML document:

<EVENT_INSTANCE>
<EventType>CREATE_TABLE</EventType>
<PostTime>2007-07-16T21:56:37.727</PostTime>
<SPID>52</SPID>
<ServerName>NETSOFT-DW</ServerName>
<LoginName>sa</LoginName>
<UserName>sa</UserName>
<DatabaseName>CLR_TEST</DatabaseName>
<SchemaName >dbo </S chemaName >
<0bjectName>TableTest</Obj ectName>
<0bj ectType>TABLE</Obj ectType>
<TSQLCommand>
<SetOptions ANSI_NULLS="ON" ANSI_NULL_DEFAULT="ON" 
	    ANSI_PADDING="ON" QUOTED_IDENTIFIER="ON" ENCRYPTED="FALSE" />
<CommandText>create table TableTest(id int)</CommandText>
</TSQLCommand>
</EVENT_INSTANCE>

You can use XQuery methods to parse the content of an event document:

CREATE TRIGGER DDL_ProtectTable_Table

ON DATABASE

AFTER DDL_TABLE_EVENTS

declare @event xml

set @event = EVENTDATA()

INSERT INTO [dbo].[ActivityLog]([Activity],[LogDate],[UserName],[Note])

VALUES

(@event.value('(/EVENT_INSTANCE/EventType)[1]', 'nvarchar(100)'),

GETDATE(),

@event.value('(/EVENT_INSTANCE/LoginName)[1]', 'nvarchar(100)'),

@event.value('(/EVENT_INSTANCE/TSQLCommand)[1]', 'nvarchar(2000)') );

DDL triggers

Mon, 16 Jul 2007 09:55:41 GMT

DDL triggers are new in SQL Server 2005. They can be fired after the engine invokes Data Definition Language (DDL) statements such as:

Create Table
Drop Table
Alter Procedure
Drop Schema
Create Login

DDL triggers are typically used to audit or control changes in database and server objects. In the following example, the trigger will prevent changes to any table in the current database:

CREATE TRIGGER DDL_ProtectTable_Table

ON DATABASE

FOR DROP_TABLE, ALTER_TABLE, CREATE_TABLE

RAISERROR('This database contains sensitive data. Changes to the

tables are typically not allowed and are rolled back by a trigger.', 16, 1)

ROLLBACK;

DDL triggers are After triggers, which means that:

You can replace the For keyword with After (as in the third line of the preceding example).
You have to explicitly invoke Rollback statements to cancel the changes made by DDL statements.

You can find a great example of DDL triggers on polish Technet sites: http://www.microsoft.com/poland/technet/article/art0060_01.mspx

This artcicle was written in cooperation by two polish SQL Server gurus so it is really worth of reading!!!!!

Indexing XML for performance

Sun, 08 Jul 2007 21:26:59 GMT

Sometimes you want to increase query performance by creating an index on your XML columns. SQL Server supports both primary and secondary indexes on XML columns. SQL Server will index all the tags, values, and paths in reverse
order, as path expression searches are faster when the suffix is known.
SQL Server stores the XML datatype in a BLOB field. This field can be up to 2GB in size, and
parsing this field at runtime to perform queries without an index can be very time consuming. It is important to remember that the
base table must have a clustered index to create an XML index on an XML column. You cannot
modify the primary key of the table until you drop all XML indexes on the table. If you :

- do a lot of queries but few inserts on your table with the XML column

- XML data is large but you’re often returning only a subset of the data

you should consider indexing the column with one or more indexes.

Creating the primary index is very straightforward. You can have only one primary XML index
per XML column. You can, however, have multiple secondary indexes on an XML column. I will show you examples how it works later.

SQL Server Internals - Schedulers

Sat, 07 Jul 2007 23:56:38 GMT

In SQL Server 2005, each CPU (whether hyperthreaded or physical) has its own scheduler created for it when SQL Server starts up. This is true even if the affinity mask option has been configured so that SQL Server is set to not use all of the available physical CPUs.Each scheduler is set to either ONLINE or OFFLINE based on the affinity mask settings. By default, all schedulers are set to ONLINE. You change then affinity mask value (and change the scheduler status to OFFLINE) without having to restart SQL Server. When a scheduler is switched from ONLINE to OFFLINE then any work already assigned to the scheduler is first completed and no new work is assigned