Dimant DataBase Solutions

Must read that blog

2012-01-10T20:57:00.000-08:00

Happy New Year to everyone!

Just found great web site about SQL Server maintained by Remus Rusanu (MS Employee)
It contains great info not only about published releases (SQL Server 2005/2008/2008R2) but also new features,command about SQL Server 2012

http://rusanu.com/2011/08/05/

Implicit conversions , sometimes it is hidden

2011-12-07T02:02:00.000-08:00

Consider simple table with one column defined as REAL datatype
CREATE TABLE #t (c REAL)
INSERT INTO #t VALUES (0)

SELECT COUNT(*) FROM #t WHERE c=''

In above statement you expect getting 0 rows to be returns as we filter out for all nonempty rows..But it returns 1 and the answer you find looking at execution plan.

SQL Server will implicitly convert '' to REAL datatype with 0 and a result is 1 row to be return.

Filter out characters---conversion error?

2011-11-06T06:13:00.000-08:00

Just help out may colleague to write a query where we needed to filter out all data that contains characters and because a column is defined as VARCHAR we CAST it to INTEGER in order to implement range searching. Please see simplified demo script.

CREATE TABLE #t (c varchar(50))
INSERT INTO #t VALUES ('122')
INSERT INTO #t VALUES ('4545')
INSERT INTO #t VALUES ('4545/454')
INSERT INTO #t VALUES ('4899')

----Failed
SELECT * FROM
(
SELECT c FROM #t WHERE c NOT LIKE '%[/]%'
) AS d WHERE CAST(c AS INT)>10

----Succeed
SELECT * FROM #t
WHERE CASE WHEN c LIKE '%[^0-9]%' THEN 0
WHEN CAST(c AS int) BETWEEN 1 AND 1000 THEN 1
ELSE 0 END = 1

The answer why the first attemp is failed we found looking at execution plan.
Predicate
CONVERT(int,[tempdb].[dbo].[#t].[c],0)>(10) AND NOT [tempdb].[dbo].[#t].[c] like '%[/]%'

Thinking that we filter out all "bad" rows and can CAST the rest is wrong because as we see above predicate is applied for the whole table.

As opposite the second query we used CASE expression to filter out "bad" rows CASE...=1 we see that SQL Server really filters out "bad" rows and now CAST is working.
Predicate
CASE WHEN [tempdb].[dbo].[#t].[c] like '%[^0-9]%' THEN (0) ELSE CASE WHEN CONVERT(int,[tempdb].[dbo].[#t].[c],0)>=(1) AND CONVERT(int,[tempdb].[dbo].[#t].[c],0)<=(1000) THEN (1) ELSE (0) END END=(1)

Why all my stored procedures are saved in master database under System stored procedure folder?

2011-11-03T01:00:00.000-07:00

Just having a discussion with a colleague , she made some changes in configuration and now when she creates a simple (not a system) stored procedure in master database it saves under Programmability --Stored Procedures--System Stored Procedures. Is it comfortable? No,right? After some investigation I found that we need to return 'allow updates' to 0 , see below script

EXEC sp_configure 'allow updates',0
reconfigure

Now everything got back to work in the 'right' place.

Default Schema for Windows Group

2011-09-05T02:04:00.000-07:00

Remember if you connect to SQL Server using Windows group, you cannot define a default schema. It is greyed out. This means that if you want everyone to be in a specific schema, you can't do it by default.

See reference
http://connect.microsoft.com/SQLServer/feedback/details/328585/default-schema-for-windows-group.

I hope finally MS will be addressing this issue in the next release of SQL Server named Denali (CTP 4)

/*
I'm happy to report that the next CTP for SQL Server Denali will allow default schemas to be assigned to Windows groups.

I'd like to thank everyone for their votes and attention to this issue. It made a different in being able to appropriately prioritize and make this improvement happen.

Il-Sung.
*/

SQL Server 2008 R2 Service Pack 1 has been released

2011-07-12T20:58:00.000-07:00

MS has just released SQL Server 2008 R2 Service Pack 1. Both the Service Pack and Feature Pack updates are available for download on the Microsoft Download Center. SQL Server 2008 R2 SP1 contains cumulative updates for SQL Server 2008 R2 and fixes to issues that have been reported through our customer feedback platforms.

To download the Service Pack packages please visit:

http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=26727 ---SQL Server 2008 R2 SP1
http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=26729 ---SQL Server 2008 R2 SP1 Express
http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=26728 ---SQL Server 2008 R2 SP1 Feature Packs

To learn more about SQL Server 2008 R2 SP1 please visit Knowledge Base Article For Microsoft SQL Server 2008 R2 SP1.
http://support.microsoft.com/kb/2528583

Partitioning experience

2011-06-13T02:15:00.000-07:00

As you know since SQL Server 2005 MS introduced partitioning feature. I have been using it frequently with my clients but yesterday ,answering question on MSDN forum about partition I learned something new. Let say we have a partition table and staging table with the data and check constraint. See the below script.

create database db1
go
use db1
go

create partition function pf1(int) as range left for values (2,4)
create partition scheme ps1 as partition pf1 all to ([primary])

create table db1.dbo.t1 (id int ,descr varchar(2)) on ps1(id)
create table db1.dbo.t2 (id int,descr varchar(2))
alter table db1.dbo.t2 add check (id > 9)

create nonclustered index i1 on db1.dbo.t1(id)
create nonclustered index i2 on db1.dbo.t2(id)

insert into db1.dbo.t1 values (1,'n1')
insert into db1.dbo.t1 values(2,'n2')
insert into db1.dbo.t1 values(3,'N3')
insert into db1.dbo.t2 values (10,'n1')
insert into db1.dbo.t2 values(15,'n2')

---Add new range
alter partition function pf1() split range (9)

---Moving the data from t2 to t1 into a new added range
alter table db1.dbo.t2 switch to db1.dbo.t1 partition 4

Msg 4972, Level 16, State 1, Line 2
ALTER TABLE SWITCH statement failed. Check constraints or partition function of source table 'db1.dbo.t2' allows values that are not allowed by check constraints or partition function on target table 'db1.dbo.t1'.

I got the error, but everything should be ok. Both tables have the same datatypes,indexes,lengths.. What is going on here. I spend some time and found the below explanation in the BOL

/*
We recommend that you specify NOT NULL on the partitioning column of partitioned tables, and also nonpartitioned tables that are sources or targets for ALTER TABLE...SWITCH operations. By doing this, you make sure that any CHECK constraints on partitioning columns do not have to check for null values. Null values are ordinarily placed in the leftmost partition of a partitioned table. When you are switching any partition other than the leftmost partition and when the ANSI_NULLS database option is set to ON, the absence of a NOT NULL constraint on the source and target tables might interfere with any CHECK constraints that are also defined on the partitioning column.
*/
So altering CHECK constraint with id IS NOT NULL solved the problem..

alter table db1.dbo.t2 add check (id > 9 AND id is not null)

SQL Server 2005 Maintenance Plan won't delete bakup files created in SQL Server 2008

2011-05-03T02:52:00.001-07:00

I have been asked recently to help to one of my friends with "strange behaviour" as he described...He has SQL Server 2008 where sql job copies .BAK files (database backup) into his local disk drive with requirement to keep those file only for one month. So as he has SQL Server 2005 (SP3) installed on his machine the first idea was to create Maintenance Clean Up task to delete the files. But as you imagine that did not work. Surely, when I came to help, I did not know that those files are created in SQL Server 2008 and has been copied to his local machine and after cheching all possibilities I launch up SQL Server Profiler to see what is going on... I saw the below command to be executed for each file to be deleted .
exec sp_executesql N'RESTORE LABELONLY FROM DISK=@P1',N'@P1 nchar(27)',N'X:\GBDB\Log\log080411PM.bak'
Well, I put that statement in Query window and obviously got the error which leads me to the root of the problem.
Msg 3241, Level 16, State 7, Line 1
The media family on device 'X:\GBDB\Log\log040411PM.bak' is incorrectly formed. SQL Server cannot process this media family.
Msg 3013, Level 16, State 1, Line 1
RESTORE LABELONLY is terminating abnormally.

What do you think guys? Does SQL Server need to check out the version of the files need to be deleted or just delete files with .BAK extension regardless on version they were created ?

Name of CTE and referenced table should be different

2011-04-06T00:20:00.001-07:00

Just got the below example from the public forum.

CREATE TABLE Batch (Batch CHAR(1),Status INT)

INSERT INTO Batch VALUES ('A',1)
INSERT INTO Batch VALUES ('B',2)
INSERT INTO Batch VALUES ('C',3)

WITH Batch AS
(
SELECT *, ROW_NUMBER() OVER(ORDER BY Batch, Status) AS RowNum
FROM Batch
)
DELETE FROM Batch
WHERE RowNum=1

As you can see , running the query SQL Server throws the error.
Msg 252, Level 16, State 1, Line 1
Recursive common table expression 'Batch' does not contain a top-level UNION ALL operator.

SQL Server "thinks" that the CTE referenced to itself but there is no UNION ALL clause.In the blow example CTE named EmpCTE referenced within to itself to join with Employees table.

WITH EmpCTE(empid, empname, mgrid, lvl)
AS
(

-- Anchor Member (AM)
SELECT empid, empname, mgrid, 0
FROM Employees
WHERE empid = 7
UNION ALL

-- Recursive Member (RM)
SELECT E.empid, E.empname, E.mgrid, M.lvl+1
FROM Employees AS E
JOIN EmpCTE AS M
ON E.mgrid = M.empid
)
SELECT * FROM EmpCTE

Finally,in order to resolve the problem you need that CTE and user table have different names. Something like that

WITH Batch_cte AS
(
SELECT *, ROW_NUMBER() OVER(ORDER BY Batch, Status) AS RowNum
FROM Batch
)
DELETE FROM Batch_cte
WHERE RowNum=1

MVP Award for 2011 year.

2011-04-04T01:55:00.000-07:00

I just got my MVP renewal email from MS, so this is my 5th award for SQL MVP since 2007.
I would say thanks for all support and good communication with all MVP around the world and hopefully it could make me more motivated to give more contribution for the community

To someone who specializes in SQL Server performance tuning

2011-03-30T03:27:00.001-07:00

Must read Conor's blog
http://blogs.msdn.com/b/conor_cunningham_msft/

Alias issue in T-SQL or defensive programming

2011-02-06T00:39:00.000-08:00

Recently I have talked to our developer who wanted to delete TOP x rows from the table. I pointed him to the below artcile http://blogs.msdn.com/b/sqlcat/archive/2009/05/21/fast-ordered-delete.aspx where a tip – a view with ORDER BY.

As alternative he wanted using a derived table but cannot understand why all rows are deleted from the table instead of TOP(x). See the below demo.

create table #t (c int)
insert into #t values (1)
insert into #t values (1)
insert into #t values (2)
insert into #t values (3)
insert into #t values (3)

delete #t from (select top (2) c
from t order by c) t

How does DELETE extension in T-SQL work?. The FROM clause after the DELETE specifies the target table to delete. The second optional FROM clause specifies the qualifying rows. But if I change 't' alias to '#t' as original name of the temporary table that would work...

delete #t from (select top (2) c
from t order by c) #t

Now, SQL Server 'sees' that derived table has the same name as a target and thus deletes only TOP(x) rows

There is no goal of this post to get into a discussion about how to write correlated subquery to perform such operations, I just wanted you to pay attention on if you choose using derived tables to perform deletion please make sure that alias you specify for derived table is the same as a target table..

PS. If you are testing and not sure about the result please use BEGIN TRAN... before executing the script.If you see that rows affected by the script is too many issue ROLLBACK TRAN to back to original data.

Dedup on huge table

2011-01-11T22:57:00.000-08:00

Hi friends
At this time I would like to share with you my experience to delete duplicates in very large table(800 million of rows).
A very general method is to use ROW_NUMBER function to PARTITION ON desired columns and then filter out only unique data.

WITH cte
AS
(
SELECT ,ROW_NUMBER() OVER (PARTITION BY ORDER BY ) rn
FROM tbl
) SELECT * FROM cte WHERE rn=1

As you imagine on huge table it will take too long. In order to optimize that query I used batch processing (divided that transaction into small chunks)

DECLARE @x INT
SET @x = 1
WHILE @x < 44,000,000 -- Set appropriately
BEGIN

;WITH cte
AS
(
SELECT ,ROW_NUMBER() OVER (PARTITION BY ORDER BY ) rn
FROM tbl WHERE ID BETWEEN @x AND @x + 10000
)SELECT * FROM cte WHERE rn=1

SET @x = @x + 10000
END
Ok ,it worked much faster as we have a clustered index on ID column such as SQL Server uses it to get the data based on defined range.However, we have another problem that data we are getting back is not actually unique.You see that for specific range ,say (from 1 to 10000) I get the data based on required partition and filter out for rn=1, BUT it is possible that the same row will occur in the next chunk (from 10001 to 20000) and we will also get it back because SQL Server does not recognize it as duplicate we have already got from the first chunk.

More reliable solution is checking on entire table and not to base on ranges.
SELECT ,COUNT(*) rn FROM
tbl GROUP BY
HAVING COUNT(*)>1

As you can see it could take long time ,so I also tried to create a Dedupt table with a key IGNORE_DUP_KEY option (thanks to Hugo) but insert into the table was pretty slow as well. Peter Larsson a fellow MVP has suggested the below technique that worked pretty well

CREATE TABLEe #unique (id primary key clustered)
INSERT INTO #temp (id) SELECT MIN(ID) AS ID
FROM tbl
GROUP BY Col1, col2, col3... (here you decide the uniqueness)

--And then insert into batches for

SET @id = 0
WHILE @id < (800 million or more)
BEGIN
SELECT t1.ID, t1.Col
FROM dbo.Table1 AS T1
INNER JOIN #unique as u ONu.id = t1.id
AND u.id BETWEEN @id AND @id +99999

SET @id += 100000
END

It would be great it put back here testing results.

What business says does not mean what businees wants

2010-12-16T02:28:00.001-08:00

This great sentence I learned form my exerience being a consultant

What business says does not mean what businees wants

TechEd 2010 in Eilat

2010-11-23T01:19:00.001-08:00

Hi
I am going to attend TechEd 2010 in Eilat next week. It is great opportunity to learn new things , meet new and old friends.It is my second TechEd and I am will be focusing on Data Platform direction and BI. Hope to see you there.

http://www.microsoft.com/israel/TechEd2010/Tracks/BI.aspx

Blobs and covering indexes

2010-11-08T23:26:00.001-08:00

Nowadays it is common to store/save pictures,documents in the databbase. Since SQL Server 2005 we use VARBINARY(MAX) datatype to store such data. The 'problem' I have seen recently at the client database is SELECT statement on very huge table that contain BLOB data and that data needs to be return to the client works pretty slowly. Run the below
SET STATISTICS IO ON

SELECT col1,col2, blobdata FROM tbl WHERE....

Table 'tbl'. Scan count 1, logical reads 8125, physical reads 0, read-ahead reads 0, lob logical reads 261521, lob physical reads 0, lob read-ahead reads 0.
CPU time = 1252 ms, elapsed time = 12632 ms.

We see high number of reads to return the data from index/blob pages.

Execution plan shows that there is Bookmark to return the BLOB data from Clustered Index Key because of our NCI(NonClustered Index) does not cover all columns in SELECT statement.
Ok, at first glance you would re-create NCI to INCLUDE blobdata column to 'cover' SELECT ...Well, I noticed that recreating NCI index takes long long time and LOG file was grown dramatically. I see that after INCLUDE blobdata column I reduced logical reads as shown below

Table 'tbl'. Scan count 1, logical reads 14, physical reads 0, read-ahead reads 0, lob logical reads 421521, lob physical reads 0, lob read-ahead reads 0.
CPU time = 1622 ms, elapsed time = 16512 ms.

But what happened? BLOB page reads were done almost 1.8 times more with covering index and actually I have not seen much improvment in performance of the query.
So finally, if you create an index or INCLUDE the BLOBs SQL Server create copy of that index for every blob column and thus it takes time to create index and reading the data from index page. Well, you were able to save some IO by covering BLOB column but internally SQL Server works much hardly to return the data and maintain the index.

Sometimes it is important to choose right collation for the server

2010-10-03T05:39:00.001-07:00

The title looks like a joke, am I right? Sometimes we just take care of collation for specific database. Well that may lead you to unpleasant situation and moreover getting wrong or no data at all. Recently I helped my friend to solve the problem for his new software. He created a database with "right" collation let say Hebrew_CI_AS and the program searches for the data that contains hebrew characters. Suddenly the program does not return the data at all. He showed me that the data he is looking for are presented in the table and point me to the stored procedure that performs the searching. You have already guessed what was the problem, right? The problem was because withing a stored procedure he used a temporary table where the programm stores the data for final result (VARCHAR(n) for search string) for some logic. As you know tempdb is a system database which inherits a server's collation (SQL_Latin1_General_CP1_CI_AS) and hence no data were returned. In that specific case we changed the data type to NVARCHAR(n) and there is no need to reinstall the entire server, but think about the program(I have seen many of such) that such technique (store unicode character in temporary tables) are used in hundred of stored procedures and change the datatype for all of them is very big headache.

Window Functions (OVER Clause) – Help Make a Difference

2010-09-19T02:36:00.001-07:00

Hello friends

As you probably know those days SQL Server 2011 is being developed. Itzik Ben-Gan believes (me too) that having those new window functions and enhance existing ones will benefit all of us SQL Server developers in the future more complete support for the standard window functions in SQL Server. Please read and vote..

http://www.sqlmag.com/blogs/puzzled-by-t-sql/tabid/1023/entryid/13085/Window-Functions-OVER-Clause-Help-Make-a-Difference.aspx

Be careful to grant dbCreator server role to the user

2010-09-02T02:22:00.001-07:00

It is common that vendors ask for permission to create databases (or they applications need) on your servers and most of DBAs I have seen immediately grant them dbCreator server role. But they are not aware that members of that role are able to DROP/ALTER any databases on the entire server regardless of whether or not you even have a user account in the database.Did you really want that??

The right approach is to grant CREATE ANY DATABASE permission and then the user is able to DROP/ALTER he/she owns.

Another way to execute stored procedure

2010-08-18T21:20:00.001-07:00

Just "discovered" that if you right click on the stored procedure you can clcik "Execute Stored Procedure" option. It gives you friendly GUI to provide parameters,however,as a data professional, I prefer to execute stored procedures in query builder. But it is nice feature for less experienced users.

Find dependency task again

2010-07-13T04:07:00.001-07:00

If you have ever dealt with projects where you need to script out the objects and the recreate them on the destination database you first always have been thinking of sys.sql_dependencies system table (if you are on SQL Server 2005 and onwards) . At our company we have a project to create script on many client machines that have variuos versions of SQL Server Express Edition from SQL Server 2000 to SQL Server 2008 R2. The team has automated script that creates all objects on the client by using old syscomments table. They could not rely on sys.dependencies table because it has some critical bugs as you are aware of. They did script out all views into a text file and run it for more than once because it is possoble that for the first run it is failed (if you have nested views.).It is very complicated script and in narrow down we can reproduce the problem as the following.

USE AdventureWorks
GO
CREATE TABLE t1 (c INT)
GO
CREATE VIEW v1
AS
SELECT c FROM v2
GO
/*

Msg 208, Level 16, State 1, Procedure v1, Line 3
Invalid object name 'v2'
*/
CREATE VIEW v2
AS
SELECT * FROM t1
--Clean up
DROP VIEW v1,v2
DROP TABLE t1

It worked well for a while but very slowely and they look for improvment. I would like to share with you very reliable and fast option developed by colleague of mine by using VB.NET. Please test it before running on production server.

This script reads the view definition, find the level of dependecy, and then set priority to create.

Var
_dsV – Data table of views and column of “PL” – Place or priority
_dsD - data table of dependences (view on view )
_dsC –context of views

Function
CreateDependence(_dsV.Tables(0), _dsC.Tables(0), _dsD) - find on context of view if exist other views
CheckChildViews(_dsV.Tables(0), _dsD, _dsD.Select(), 0) - recursive function which set priority of views.
On the start all view with priority 1000 (it’s don’t have dependence) . When view has some dependence then it gets priority = 0 or parent priority + 1. All child views (nested view) get higher among (parent priority +1) or its priority.
GetViews(_dsV.Tables(0), _dsC.Tables(0)) - get all view by sort of priority descending

Private Sub CreateViews()
Dim cldb As clDB ‘ class - connection to database
Dim _dsV As DataSet
Dim _dsD As New DataTable("DEP")
Dim _dsC As DataSet
lblResultSqlScript.Text = ""
Try
cldb = New clDB
cldb.ConnectionString = conectionString
Dim tmp As String = ""
'all views
tmp = " SELECT ID,NAME,1000 as PL "
tmp &= " FROM dbo.sysobjects "
tmp &= " WHERE xtype = 'V' and category=0 "
tmp &= " order by name " & vbNewLine
_dsV = cldb.GetDataSet(tmp)

'context of views
tmp = " SELECT text ,COLID,ID "
tmp &= " FROM syscomments "
tmp &= " ORDER BY COLID " & vbNewLine
_dsC = cldb.GetDataSet(tmp)

'depandences
CreateDependence(_dsV.Tables(0), _dsC.Tables(0), _dsD)

ClearIndex(_dsV.Tables(0), _dsD)
CheckChildViews(_dsV.Tables(0), _dsD, _dsD.Select(), 0)
-----GetViews(_dsV.Tables(0), _dsC.Tables(0))
RunViews (_dsV.Tables(0), _dsC.Tables(0))
Catch ex As System.Exception
Finally
cldb = Nothing
End Try
End Sub

Private Sub CreateDependence(ByVal dsV As DataTable, ByVal dsC As DataTable, ByRef dt As DataTable)
'create datatable
Dim cl1 As New DataColumn("PNAME", System.Type.GetType("System.String"))
Dim cl2 As New DataColumn("PID", System.Type.GetType("System.Int32"))
Dim cl3 As New DataColumn("CNAME", System.Type.GetType("System.String"))
Dim cl4 As New DataColumn("CID", System.Type.GetType("System.Int32"))
dt.Columns.Add(cl1)
dt.Columns.Add(cl2)
dt.Columns.Add(cl3)
dt.Columns.Add(cl4)
Dim cur As Integer = -1
Dim tmp As String = ""

'load view
Dim bs As New Text.StringBuilder
For Each ddr As DataRow In dsV.Select("", "PL DESC")
Try
bs.Length = 0
For Each dr As DataRow In dsC.Select("ID=" & ddr("ID"), "COLID")
Try
bs.Append(dr("TEXT").ToString)
Catch ex As System.Exception
End Try
Next
'find dependence
'loop on all view without current
For Each ddr1 As DataRow In dsV.Select("ID<>" & ddr("ID"), "PL DESC")
cur = bs.ToString.IndexOf(ddr1("NAME").ToString)
If cur > -1 Then
'check if it's real name and not peace of name
tmp = bs.ToString.Substring(cur + ddr1("NAME").ToString.Length, 1)
If tmp = "" Or tmp = "." Or tmp = " " Or Asc(tmp) = 13 Then
'add to dependence
Dim newdr As DataRow = dt.NewRow
newdr("PNAME") = ddr("NAME")
newdr("PID") = ddr("ID")
newdr("CNAME") = ddr1("NAME")
newdr("CID") = ddr1("ID")
dt.Rows.Add(newdr)
End If
End If
Next
Catch ex As System.Exception
End Try
Next
End Sub
Private Sub GetViews(ByVal dsV As DataTable, ByVal dsC As DataTable)

txtSQLScript.Text = ""
Dim bs As New Text.StringBuilder
bs.Append(" SET QUOTED_IDENTIFIER OFF " & vbNewLine)
bs.Append(" GO " & vbNewLine)
bs.Append(" SET ANSI_NULLS OFF " & vbNewLine)
bs.Append(" GO " & vbNewLine)
For Each ddr As DataRow In dsV.Select("", "PL DESC")
bs.Append(" IF EXISTS (select * from dbo.sysobjects where id = object_id(N'" & ddr("NAME") & "') and OBJECTPROPERTY(id, N'IsView') = 1)" & vbNewLine)
bs.Append(" DROP VIEW " & ddr("NAME") & vbNewLine)
bs.Append(" GO " & vbNewLine)

For Each dr As DataRow In dsC.Select("ID=" & ddr("ID"))
Try
bs.Append(dr("TEXT").ToString)

Catch ex As System.Exception
End Try
Next
bs.Append(vbNewLine & " GO " & vbNewLine)
bs.Append(vbNewLine & "--------------------------------" & vbNewLine)
Next
txtSQLScript.Text = bs.ToString
End Sub
Private Sub ClearIndex(ByRef dsV As DataTable, ByVal dsd As DataTable)
For Each ddr As DataRow In dsV.select
If dsd.Select("PID=" & ddr("ID").ToString).Length > 0 Then
ddr("PL") = 0
End If
Next
End Sub

Private Sub CheckChildViews(ByRef dsV As DataTable, _
ByVal dsd As DataTable, _
ByVal dr() As DataRow, ByVal index As Integer)
For Each ddr As DataRow In dr
Try
Dim pd As DataRow = dsV.Select("ID=" & ddr("PID").ToString)(0)
If pd("PL") < index Then pd("PL") = index
CheckChildViews(dsV, dsd, dsd.Select("PID=" & ddr("CID").ToString), index + 1)
Catch ex As System.Exception
End Try
Next
End Sub

Private Sub RunViews(ByVal dsV As DataTable, ByVal dsC As DataTable)
txtSQLScript.Text = ""
Dim bs As New Text.StringBuilder
Dim cldb As New dbAdministration.clDB
For Each ddr As DataRow In dsV.Select("", "PL DESC")
bs.Length = 0
bs.Append(" IF EXISTS (select * from dbo.sysobjects where id = object_id(N'" & ddr("NAME") & "') and OBJECTPROPERTY(id, N'IsView') = 1)" & vbNewLine)
bs.Append(" DROP VIEW " & ddr("NAME") & vbNewLine)
cldb.Execute(bs.ToString()) 'run delete if exist view

bs.Length = 0

For Each dr As DataRow In dsC.Select("ID=" & ddr("ID"))
Try
bs.Append(dr("TEXT").ToString)
Catch ex As System.Exception
End Try
Next
cldb.Execute(bs.ToString()) 'run create view
Next
End Sub

Column level permission

2010-06-08T04:52:00.000-07:00

Uh,it's so hard to back to work from the vacation:-).Yesterday,for the first time, I have been faced the requirement to implement a column level permission on the table, have you done something before? A little script below demonstrates it.

USE dbname;
CREATE TABLE demo (c INT, c2 INT)

GRANT SELECT (c) ON demo TO test

EXECUTE AS USER = 'test';

SELECT * FROM fn_my_permissions('dbo.demo', 'OBJECT')
ORDER BY subentity_name, permission_name ;

REVERT

PS: subentity_name tells you on what column given permission.

It's The Time To Go on Vacation

2010-05-16T05:46:00.001-07:00

Yep, on 21st May I am going to Moscow to see a World Table Tennis Championship. I love playing table tennis and even took a part of two World Table Tennis Championships in earliest 90. So I hope to meet there may old friends from all around the world, so I am very exited about it.. See you later on.

Kalen Delaney's seminar

2010-05-07T00:53:00.000-07:00

Just backed home from the three days of seminar in Tel-Aviv named "Deep dive into SQL Server Internals with Kalen Delaney".
Two days Kalen explained about SQL Server Query Plans,Tuning and Recompile techniques and the last one was about SQL Server Data Storage Formats.

You know , I thought I know a lot about SQL Server performnace tunning and etc, however, attending that seminar I just realized how much I need to learn. There are lot if tips,thoughts how to tune the system, indexes and etc.

I would also mention about how easy was to understand Kalen (she spoke english at seminar :-)) and her methods to provide courses which is so important for attendants.

Don't speak to the DBA while he/she is doing the job.

2010-04-12T04:25:00.002-07:00

Our DBA was very busy today, helped out our developers top write efficient code,explained to the programmer how to launch the old DTS package on SQL Server 2005 and etc. You know how it is, someone comes in and asking the question and you answer the question at the same time you are doing the work.
So,today answering one such questions our DBA deleted by mistake very ctitical database. Fortunately , we had zero data loss thanks to our backup/recovery strategy.

I think that is acceptable to say that you are busy right now please ask the question later on, what do you think?

Getting column level permision report

2010-03-16T00:34:00.002-07:00

If you think about giving column level permission , you'll probably create a view with only needed columns to expose to ,then grant SELECT permission on that view and absolutely DENY access to underlaying tables to the user, right? Also , it depends on the policy of the company but you are able to grant SELECT/UPDATE/DELETE/INSERT operations on column level, have you ever used it:-)? So I recently visited our client who does perfom security permission on column level and wanted to know what are the columns and on what tables that specific user has access? I decided to perform the below testing, see if that might help you.

CREATE TABLE demo (c INT, c2 INT)

/* Grant SELECT permission on column named c on demo table*/
GRANT SELECT (c) ON demo TO test

EXECUTE AS USER = 'test';

SELECT * FROM fn_my_permissions('dbo.demo', 'OBJECT')
ORDER BY subentity_name, permission_name ;

REVERT

Have you dealt with ENTER key in INSERT table?

2010-03-03T02:33:00.002-08:00

I have a client who uses the application that allows to insert data but also typing ENTER key in the middle. Well, the question is how do you want to store the data and later displaying on the client? He was not sure , so I showed him the below technique to display such kind of data....

--Create auxiliary table
SELECT TOP 8000 n = IDENTITY(INT)
INTO Numbers
FROM syscolumns s1, syscolumns s2 ;

CREATE TABLE test (col1 varchar(8000))

INSERT INTO test values ( 'Row----------1
Row--------2
Row--3
Row----4'
)

---See how it stores the data.
SELECT * FROM TEST

--Remove ENTER key
SELECT SUBSTRING(col1, n, CHARINDEX(CHAR(13), col1 + CHAR(13), n) - n)
FROM Numbers, test
WHERE SUBSTRING(CHAR(13) + col1, n, 1) = CHAR(13)
AND n < LEN(col1) + 1 ;

--Cleanup
DROP TABLE Numbers
DROP TABLE test

I was invited to blog on www.sqlblog.com

2010-02-03T04:53:00.002-08:00

A few days ago Adam Machanic who is the owner of that great web site invited me to join his great company. I am really excited about and proud to be part of such great company of friends who write very good stuff about SQL Server.
I would like to keep bloging here as well , I do not know how long it will take but you can always catch me up on http://sqlblog.com/blogs/uri_dimant/

Think before unchecking sysadmin rights of BUILTIN\Administrators.

2010-01-12T03:53:00.000-08:00

I have recently met our client who uchecked the sysadmin rights of BUILTIN\Administrators group before given any permissions to anotgher account.

That was NOT such problem if the BUILTIN\Administrators group was removed from sysadmin role accidentally/by mistake, then you must login with another sysadmin login. If there is no other sysadmin login, you must login with SQL authentication as sa with the password that was set during setup to sa. Once logged in as a member of sysadmin, you are able to add BUILTIN\Admisnitrators back to sysadmin role.
However everything above does not work for the client. Uhhh,the client also disabled SA accoount as well as DAC connection.
Moreover, there is no domain controller where you can create a sysadmin domain acoount and grant the access to the machine running SQL Server,that was a stand alone computer with single instance installed on.

The solution we found was to start SQL Server with single user mode. Using the single-user mode, SQL Server 2005 prevents a Windows Administrator to abuse this privilege to act on behalf of the sysadmin without being noticed. This allows Windows Administrator accounts to perform certain maintenance tasks, such as installing patches. To someone who is not familiar how to start the instance in single user mode and adding login to the server role being system administrator please read the below link describing step by step the procedure.

http://blogs.msdn.com/raulga/archive/2007/07/12/disaster-recovery-what-to-do-when-the-sa-account-password-is-lost-in-sql-server-2005.aspx

SQL_VARIANT_PROPERTY() to find the info about data type?

2009-12-20T07:21:00.000-08:00

BOL says that this function returns the base data type and other information about a sql_variant value. Have you ever looked what is datatype of GETDATE() or perhaps DB_ID() function or what is the data type of SYSTEM_USER??

DECLARE @var sql_variant
SET @var =GETDATE() --SYSTEM_USER
SELECT SQL_VARIANT_PROPERTY(@var , 'BaseType') as BaseType
, SQL_VARIANT_PROPERTY(@var , 'Precision') as Precision
, SQL_VARIANT_PROPERTY(@var , 'Scale') as Scale
, SQL_VARIANT_PROPERTY(@var , 'TotalBytes') as TotalBytes
, SQL_VARIANT_PROPERTY(@var , 'MaxLength') as MaxLength
, SQL_VARIANT_PROPERTY(@var , 'Collation') as Collation;

I found that very usefull as you can easily retun the basic info about even system objects and how SQL Server interpetes it.

Think before you use ALTER TABLE ....

2009-12-09T01:14:00.000-08:00

Let me say you have been told to add a computed column to the very huge table in SQL Server. Well,if you have been working with SQL Server for a while you are probably aware that you will be better of using ALTER TABLE.. command instead of using SSMS because under some circumstances SQL Server will copy the entire table into a temporary one and then rename it , but if the users work on the table they will be locked... So far so good. Now , what if you have been told to add a computed column with PERSISTED clause which means SQL Server will store physically the values of the column..Here stop and think a little bit especialy you deal with huge tables. SQL Server is going to update every row in the table with computed value based on the formula you provide with. How long will it take? I have seen that adding a PERSISTED column by using ALTER TABLE command on the table with 230 million rows took 19 hours on very powerful server, can you afford it?.
So what is the solution? The solution is to add permanent column and then running an UPDATE command , however do not forget to divide the UPDATE into small batches (each one of 10000 rows for example) such as you will control the rows has been updated/affected by the UPDATE and your LOG file will not be blowing up.

Consistency checking for a 1.2 TB database

2009-11-08T06:11:00.000-08:00

I would like to share with you my recent experience on how to build consistency checking for VLDB by using DBCC CHECKDB/CHECKTABLE and etc. I was following Paul's great article
http://www.sqlskills.com/BLOGS/PAUL/post/CHECKDB-From-Every-Angle-Consistency-Checking-Options-for-a-VLDB.aspx
to build the same or almost the same strategy on the client's site. They have 8 huge tables arround 658,245,225 pages each table and I tested the bigest one to run DBCC CHECKTABLE command. It took 17 hours to complete the task but unfortunately it was not acceptable for the client. So my final solution is (thanks to Paul) to restore the database on the dev.machine and perform the checking on.

Import data from AS400 by using SSIS

2009-10-18T05:37:00.000-07:00

That was deviation towards the world of BI. I was helping my co-worker to build a SSIS Package to import the data from AS400 into SQL Server.
At first I selected OLE DB Source component for getting the data but started getting odd errors of driver compatibilty. Then I read that there is a newish driver called DB2OLEDB which unfortunately works only on Enterprise/Developer Editions ,so I had only Standtard...

After some investigations I solved the problem by choosing DataReader Source control and selected ODBC driver for Net.Providers. It looks obvious for experienced BI consultants but for me, who rarely use SSIS, I felt great when co-worker was so happy to get all the data:-))

Does the fragmentation have impact on performance on Heaps?

2009-10-05T03:32:00.000-07:00

I have been working on some project that involves using very huge tables. Actually most of tables are heaps and to reduce IO we decided to compress those tables on PAGE level because the database is READ-ONLY. Fortunately in SQL Server 2008 we can issue ALTER TABLE... REBUILD.. to rebild the heaps to remove forwarding pointers ,please read Tibor's article (http://sqlblog.com/blogs/tibor_karaszi/archive/2009/08/28/heaps-forwarding-pointers-alter-table-rebuild-and-non-clustered-indexes.aspx)

Well ,after running the scripts we ended up with almost 500GB empty space. Usually it is not good to shrink the database as you gain nothing (it will grow again) and that caused fragmentation which does not help in terms of performance. Shrinking VLDB is another topic but what I did NOT know if your queries are doing INDEX SEEK SQL Server can get to any row easily no matter how fragmented the table is.
In that case fragmenation has no impact on performance.

If you think about it does perfect sense. Fragmentation is a measure of how closely the logical order matches the physical order. So if you do a scan (following the logical order) on a fragmented table, you can be jumping all over the to get the rows you need,instead of accessing them nice and neat and contiguously.

Many thanks to Kalen Delaney.

UNION ALL within a view

2009-09-02T00:44:00.000-07:00

I have a client who has viewes that contain UNION ALL command for number of very large tables. The logic is to apply a WHERE condition to that view to get the data. Even having indexes on those tables SQL Server applies a WHERE condition to each SELECT statement within a view that may lead to performance proble. To demonstate it please considetr AdwentureWork data and two tables Sales.SalerDetails and SalesOrderHeader

CREATE VIEW v1
AS
SELECT S.SalesOrderID
FROM
Sales.SalesOrderDetail S
UNION ALL
SELECT S.SalesOrderID
,S.CreditCardID FROM
Sales.SalesOrderHeader S

SET STATISTICS IO ON
SELECT TOP 100 * FROM v1 WHERE SalesOrderDetailID >40000 AND SalesOrderDetailID<45000

Table 'SalesOrderDetail'. Scan count 3, logical reads 1359, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'SalesOrderHeader'. Scan count 1, logical reads 34, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

You can see that SQL Server 'touched' those tables by using Clustered Index Scan on
Sales.SalerDetails and Clustered Index Seek on Sales.SalesOrderHeader. Indeed I realy simplified the logic as in reality it has horrible performance. How to improve?

I create a Multi-Statement Table-Valued UDF that accepts a parameter
CTREATE FUNCTION dbo.udf1
(
@SalesOrderID INT
)
RETURNS @t TABLE (c1 INT,c2 INT)
AS
BEGIN
INSERT INTO @t
SELECT S.SalesOrderID
FROM
Sales.SalesOrderDetail S
WHERE SalesOrderID > @SalesOrderID AND SalesOrderID<@SalesOrderID+5000
UNION ALL
SELECT S.SalesOrderID
FROM
Sales.SalesOrderHeader S
WHERE SalesOrderID > @SalesOrderID AND SalesOrderID<@SalesOrderID+5000
RETURN
END
Table '#74AE54BC'. Scan count 1, logical reads 12, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

See , I applied a WHERE condition within an UDF for every statement as UDF must get parameters.This is another argument to use Table-Valued UDF ...

SQL Server 2005/2008 on Windows 7

2009-08-09T23:32:00.000-07:00

SQL Server MVP John Paul Cook has already mentioned on www.sqlblog.com that SQL Server 2008 will install on Windows 7 and it works just fine. I would like to add that SQL Server 2005 is also working just fine along with SQL Server 2008.
Having said that I'm told by many people that Reporting Services does not work on W7.Well the phrase 'does not work' does not say too much without additional info however I decided to do some testing and found out that SSRS works just fine as well, so the problem could be because the SSRS/SQL Server account does not have enough credentials to run SSRS. More over I took old reports which were deployed to our dedicated SSRS server (Windows 2003) and redeployed to Windows 7 and that also worked just fine. I'm really happy with Windows 7 which is faster than Vista and more comfortable than XP.....

Report on Indexes on FKs

2009-07-08T22:56:00.000-07:00

This question had been asked in NG yesterday.I think it is a good exercise so I take a script too get the column usage from both sides (parent/child) written by Aaron Bertrand and modified a little bit to return the table/column that does not have an index on.

WITH fk_no_indexes
AS
(
SELECT
[constraint_name] = f.[name],
[child_table] = OBJECT_NAME(f.parent_object_id),
[child_table_id]=f.parent_object_id,
[child_column] = cc.name,
[child_column_id]=cc.[column_id],
[parent_table] = OBJECT_NAME(f.referenced_object_id),
[parent_column] = pc.name
FROM
sys.foreign_keys f
INNER JOIN
(
SELECTc.[object_id],c.name,c.column_id,ic.index_id
FROM sys.columns c INNER JOIN sys.index_columns ic
ON c.[object_id] = ic.[object_id] AND c.column_id = ic.column_id
) AS pc
ON
f.key_index_id = pc.index_id
INNER JOIN sys.foreign_key_columns fkc
ON f.[object_id] = fkc.constraint_object_id
AND pc.[object_id] = fkc.referenced_object_id
AND fkc.referenced_column_id = pc.column_id
INNER JOIN sys.columns cc
ON fkc.parent_object_id = cc.[object_id]
AND fkc.parent_column_id = cc.column_id
)
SELECT [constraint_name],[child_table],[child_column],[parent_table] FROM
fk_no_indexes WHERE NOT EXISTS
(
SELECT * FROM sys.index_columns i
WHERE i.[object_id]=[child_table_id]
AND [child_column_id]=column_id
)
ORDER BY
constraint_name,
child_table;

Have you applied latest sevice pack to the client tools?

2009-06-15T05:02:00.000-07:00

Well, I have seen a huge number of developers who have not heard aboit it at all. Yes,I'm serious.Recently, I had helped out tuning SQL Server performance of the database in pretty big and to be held in respect company with more than 10 developers. One day one of them asked me about the query ,which restuns the data without the error on his machine ,however , throws the error on the server.If I remember well it was conversion error.I checked both queries and indeed on the dev.machine it runs without the error. I know there is no magic here..and as you can imagine the dev.machine has installed NO service pack at all,but Production SQL Server 2000 has latest service pack (SP4).That's all story. The 'little' difference could have made a big error witout testing as each developer has installed server and client tools on their machines.The SQL Server query optimizer is free to move/change expressions with new release of service packs or hotfix thus please make sure that you have latest service pack on your client tools and test the query as well.

How to upload/modify more than one BLOB column in single statement

2009-05-14T01:24:00.000-07:00

Recently, I have been visited our client who has one table with two columns defined as VARBINARY(MAX)to store the images. How can we insert/upade these columns with single statements? I have not used blobs too much ,so I ended up with the following

CREATE TABLE Blobs (id INT NOT NULL IDENTITY(1,1) PRIMARY KEY ,b1 VARBINARY(MAX),b2 VARBINARY(MAX))

---Single INSERT statement

INSERT INTO Blobs (b1,b2)
(SELECT *,(SELECT * FROM
Openrowset(Bulk 'N:\OnePicture.JPG', SINGLE_BLOB) AS d1)
FROM
Openrowset(Bulk 'N:\AnotherPicture.JPG', SINGLE_BLOB) AS d2)

--Update

UPDATE Blobs SET b1=(SELECT * FROM
Openrowset(Bulk 'N:\AnotherPicture.JPG', SINGLE_BLOB) AS d2),
b2=(SELECT * FROM
Openrowset(Bulk 'N:\OnePicture.JPG', SINGLE_BLOB) AS d2)
---WHERE condition put here

My wife gave birth to our third child on 4 May!

2009-05-06T00:23:00.000-07:00

A son, 3 kg 370 grams!

We have replaced ourselves and are now done.

How to eat ASYNC_NETWORK_IO ?

2009-04-19T00:54:00.000-07:00

Last time I was visiting our client the person showed me one process which was in suspended status for a long time. Using Adam's great script(http://sqlblog.com/blogs/adam_machanic/archive/2009/03/30/who-is-active-v8-40-now-with-delta-power.aspx) I identified the SELECT statement which had ASYNC_NETWORK_IO wait type. MS said that this type occurs on network writes when the task is blocked behind the network and we need to verify that the client is processing data from SQL Server. The problem was that SELECT statement retrievs more than 50.000 rows , then the application manipulates (FOR ..NEXT loop) with that data and finally only 200 rows that a client had seen at the end. Well,thanks to SQL Server 2005 we changed the statement to accept a parameter for TOP clause and this wait type has gone. We have not seen a great improvment of the application as we did not expect it but it is another way to identify poor written queries.

MVP Award for 2009 year.

2009-04-01T21:07:00.000-07:00

I've been given the MVP award for another year. I hope you have found this blog and my posts in public forums useful and that helped you out to resolve some problems.

What is a shared drive/s used by SQL Server instance?

2009-03-05T05:08:00.000-08:00

SQL Server 2005 has a great dynamic management view named sys.dm_io_cluster_shared_drives . If you have clustering and want to know what
shared disk/s which is/are used by clustered SQL Server instance run the following query.

SELECT DriveName
FROM sys.dm_io_cluster_shared_drives
ORDER BY DriveName

Running DTS without SQL Server?

2009-03-03T01:44:00.000-08:00

Seems like unreasonable request.Well, we had a client who wanted by click on icon on his Desktop to launch a DTS Package which runs on another computer.I'd like to share with you the script that friend of mine created on a workstation that did not have SQL Server installed.

MD "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"

COPY sqlresld.dll "\Program Files\Microsoft SQL Server\80\Tools\Binn"

COPY dtsffile.dll "\Program Files\Microsoft SQL Server\80\Tools\Binn"
COPY dtsffile.rll "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"
regsvr32.exe "\Program Files\Microsoft SQL Server\80\Tools\Binn\dtsffile.dll"

COPY dtspkg.dll "\Program Files\Microsoft SQL Server\80\Tools\Binn"
COPY dtspkg.rll "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"
regsvr32.exe "\Program Files\Microsoft SQL Server\80\Tools\Binn\dtspkg.dll"

COPY dtspump.dll "\Program Files\Microsoft SQL Server\80\Tools\Binn"
COPY dtspump.rll "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"
regsvr32.exe "\Program Files\Microsoft SQL Server\80\Tools\Binn\dtspump.dll"

COPY axscphst.dll "\Program Files\Microsoft SQL Server\80\Tools\Binn"
COPY axscphst.rll "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"
regsvr32.exe "\Program Files\Microsoft SQL Server\80\Tools\Binn\axscphst.dll"

COPY dtsrun.exe "\Program Files\Microsoft SQL Server\80\Tools\Binn"

COPY dtsrun.rll "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"

COPY custtask.dll "\Program Files\Microsoft SQL Server\80\Tools\Binn"
COPY custtask.rll "\Program Files\Microsoft SQL Server\80\Tools\Binn\Resources\1033"
regsvr32.exe "\Program Files\Microsoft SQL Server\80\Tools\Binn\custtask.dll"

COPY sqlunirl.dll %SYSTEMROOT%\system32

PAUSE
--------------
"\Program Files\Microsoft SQL Server\80\Tools\Binn\dtsrun.exe" /FProcessAppSalesCube.dts

PAUSE

How to render the Report to PDF/EXCEL format

2009-02-04T06:44:00.000-08:00

I had a client who wanted to get his report directly to PDF file. Previously, he used to use a Report Viewer component in application and then exporting it to the appropriate format.
In this example I render my report to the PDF file that will be located on disk C:\.
This report also accepts one parameter.The key of the module is ReportExecutionService class which contains some methods that we need to use.

Please refer to the BOL
ms-help://MS.SQLCC.v9/MS.SQLSVR.v9.en/rswsref9/html/d6ce295b-25f4-4ce3-8d1a-765d7e7d9815.htm

--VB.NET

Imports Microsoft.SqlServer.ReportingServices2005

Private Sub CreatePdfFile()
Dim Res As New Execution.ReportExecutionService
Dim params(0) As Execution.ParameterValue
Res.Credentials = System.Net.CredentialCache.DefaultCredentials
Res.Url = "http://servername/ReportServer/ReportExecution2005.asmx?wsdl"
params(0) = New Execution.ParameterValue
params(0).Name = "ParamName"
params(0).Value = Value

Dim ReportPath As String = "/Reports/ReportName"

Res.LoadReport(ReportPath, Nothing)
Dim Format As String = "PDF"
Dim devInfo As String = Nothing
Dim extension As String = Nothing
Dim mimeType As String = Nothing
Dim encoding As String = Nothing
Dim warnings() As Execution.Warning = Nothing
Dim stremIds() As String = Nothing
Dim result() As Byte = Nothing
Res.SetExecutionParameters(params, "en-us") --ParameterLanguage

result = Res.Render(Format, devInfo, extension, mimeType, encoding, warnings, stremIds)
IO.File.WriteAllBytes("C:\Report.pdf", result)
End Sub

Executing multiple script files via SQLCMD

2009-01-08T00:35:00.000-08:00

I'm sure most people either DBA or developers faced at least once a requirement to execute bunch of script files. As you know we can do it by using Script Task in SSIS, however I would like to share with you a infrequent used :r parameter in SQLCMD utilty. Let say you have three files (.sql) that do an UPDATE,INSERT and finally SELECT statements. All you need is to have a single batch file which will execute those .sql files.
Create a batch file named DML.sql and add the below code
:r "c:\myUpdate.sql"

:r "c:\myInsert.sql"

:r "c:\MySelect.sql"

Make sure that you leave free space between commands.
--Usage
EXEC master..xp_cmdshell 'SQLCMD -S Server\SQLSERVERDEV2005 -i"c:\DML.sql"'

I recommend you to read this article with more examples about how using :r parameter.
http://blogs.msdn.com/patrickgallucci/archive/2007/09/03/sqlcmd-and-the-power-of-the-little-r.aspx

Happy New Year

2008-12-26T07:08:00.000-08:00

I think that is my first blog which does not relate to SQL Server,though why not?
I wish to SQL Server Consultants lots of contracts to help out businesses in 2009, I wish to SQL Server DBAs lots of interesting work but that their databases will run smoothly,I wish to SQL Server Developers to write efficient queries and resolve challenges and to every body and their families in 2009 I wish you health ,happiness,wealth and if you look forward to a good year ahead, spread happiness with these wonderful New Year wishes.

Do you always write parameterized query?

2008-12-18T01:34:00.000-08:00

As you know SQL Server creates an execution plan for the query/stored procedure and stores it in procedure cache in order to reuse it when you run the query next time.
But more and more execution plans in memory will also hurt the performance as we do not have a control the procedure cache's size so what would you do? Lets see the below scenario I've ran on the DEV(DO NOT run on Production) machine. First of all I used DBCC FREEPROCCACHE which clears the procedure cache and causes ad hoc queries to
be recompiled. Next I ran actually the same query with different parameters.

DBCC FREEPROCCACHE

SELECT * FROM Production.WorkOrder
WHERE ProductID=522

SELECT * FROM Production.WorkOrder
WHERE ProductID=737

DECLARE @i INT
set @i = 518
SELECT *
FROM Production.WorkOrder
WHERE ProductID = @i
GO
--See how many plans SQL Server created
SELECT stats.execution_count AS exec_count,
p.size_in_bytes as [size],
[sql].[text] as [plan_text]
FROM sys.dm_exec_cached_plans p
OUTER APPLY sys.dm_exec_sql_text (p.plan_handle) sql
JOIN sys.dm_exec_query_stats stats ON stats.plan_handle = p.plan_handle

/*
1 49152 DECLARE @i INT set @i = 518 SELECT * FROM Production.WorkOrder WHERE ProductID = @i
1 24576 SELECT * FROM Production.WorkOrder WHERE ProductID=737
1 24576 SELECT * FROM Production.WorkOrder WHERE ProductID=522
*/
As you see SQL Server created THREE execution plans.
Now let's wrap the query within a stored procedure and see what will be happened.

CREATE PROCEDURE spTest
@i INT
AS
SELECT *
FROM Production.WorkOrder
WHERE ProductID = @i
GO

EXEC spTest 522
EXEC spTest 737
EXEC spTest 518

--See how many plans SQL Server created
SELECT stats.execution_count AS exec_count,
p.size_in_bytes as [size],
[sql].[text] as [plan_text]
FROM sys.dm_exec_cached_plans p
OUTER APPLY sys.dm_exec_sql_text (p.plan_handle) sql
JOIN sys.dm_exec_query_stats stats ON stats.plan_handle = p.plan_handle

/*
3 40960 CREATE PROCEDURE spTest @i INT AS SELECT * FROM Production.WorkOrder WHERE ProductID = @i
*/
Wow, you we get single execution plan that SQL Server used three times.
I also recommend you to read Tony's blog about the subject
http://sqlblogcasts.com/blogs/tonyrogerson/archive/2007/07/07/procedure-cache-tuning-sizing-from-1gbyte-to-768kbytes-increase-the-size-of-usable-data-cache.aspx

Tony writes in his blog and I'm completely agree with him.
"Surprised? It's one of the reasons DBA's keep bleating on about using stored procedures, it forces the mindset to use the procedure cache more effectively. By parameterising, but preferably using stored procedures we get plan reuse which means a) less compiles thereby reducing CPU load, b) more available pages for the data cache thereby reducing physical disk IO and c) DBA’s are always right ;)."

PS.Besides using stored procedures we can use sp_executesql with parameters as below

EXEC sp_executesql N'SELECT SUM(ProductID) AS ProductTotal
FROM Production.WorkOrder
WHERE ProductID= @ProductID', N'@ProductID INT', 722

EXEC sp_executesql N'SELECT SUM(ProductID) AS ProductTotal
FROM Production.WorkOrder
WHERE ProductID= @ProductID', N'@ProductID INT', 522

.....

GO

When constraint remains in system table even the table is no longer exists

2008-12-04T05:19:00.000-08:00

Ok, I'm talking about temporary tables and yes,it could happen to you even you think that the temporary table is no longer exists. Consider the below script where stored procedure inserts the data with wrong type which caused the error(245conversion error). As you propably know we would never name the constraints for local temporary tables (as opposite to permanent tables), because if we run the following statement from two connections we get something like that.
/*
Msg 2714, Level 16, State 4, Line 1
There is already an object named 'PK_tmp1' in the database
*/

USE master
GO
ALTER PROCEDURE spSample
AS
CREATE TABLE #tmp(
col int --PRIMARY KEY NONCLUSTERED--,
CONSTRAINT PK_tmp1 PRIMARY KEY NONCLUSTERED
(
col
)
)
BEGIN TRY
INSERT INTO #tmp SELECT 'A'
END TRY
BEGIN CATCH
SELECT ERROR_NUMBER()
END CATCH

--Usage
USE master
EXEC spSample

USE tempdb
SELECT * FROM sys.objects
WHERE OBJECT_NAME(OBJECT_ID)like '%#tmp%'

Try add DROP TABLE #tmp in BEGIN CATCH ..block and then SQL Server deletes the constraint as well.

Can I hide undesirable database in SSMS?

2008-11-12T06:34:00.000-08:00

We have discussed the issue many times and there is an opened connection to Microsoft to add this feature in the next release. Recently I was visiting our client and we tried to do something for the subject.

1) CREATE new SQL login 'John'
2) CREATE a user named 'John in master database
3) GRANT CREATE DATABASE to John
4) While impersonating John, create a database called 'demo'
5) REVOKE CREATE DATABASE permission from John
6) REVOKE VIEW ANY DATABASE permission from PUBLIC
7) Register this server as John
8) From the 'John' session, expand database tree. Now, you should see only master, tempdb, dbtest
9) GRANT VIEW ANY DATABASE to PUBLIC
10) From the 'John' session, you should see all the databases.

However, this works perfectly if the user is the owner of the database.

Track down DDL Database changes

2008-10-22T06:17:00.000-07:00

If you want to know who alters/drops/creates tables/views/stored procedures... I would like to share with you the following script. As you know DDL Triggers (introduced in SQL Server 2005) work very like the DML triggers but details of the event that fired the a trigger are available only in XML format.

Fist of all I create a table that will hold events.

CREATE TABLE [dbo].[DDL_ChangeEvents](
[ID] [int] IDENTITY(1,1) NOT NULL,
[Session_ID] [int] NOT NULL CONSTRAINT [DF_ddl_change_Session_ID] DEFAULT (@@spid),
[Session_IPAddress] [nvarchar](50) NULL,
[Insert_Date] [datetime] NOT NULL CONSTRAINT [DF_ddl_change_Insert_Date] DEFAULT (GETDATE()),
[Username] [nvarchar](100) NOT NULL CONSTRAINT [DF_DDL_change_Username] DEFAULT (CONVERT([nvarchar](100),ORIGINAL_LOGIN(),(0))),
[EventType] [nvarchar](200) NULL,
[objectName] [nvarchar](200) NULL,
[objectType] [nvarchar](200) NULL,
[sql] [nvarchar](max) NULL
) ON [PRIMARY]

It wont help if I get only SPID of the session as in many cases users get logged with only one defined login or even with 'sa'.So I need IP address of those workstations thus I added Session_IPAddress column.
Now, let's create a database trigger to capture the info.

CREATE TRIGGER [trgDataDDLChangeEvent] ON DATABASE
FOR DDL_DATABASE_LEVEL_EVENTS
AS
DECLARE @eventdata XML
SET @eventdata = EVENTDATA()
IF @eventdata.value('(/EVENT_INSTANCE/EventType)[1]', 'nvarchar(200)')
<> 'CREATE_STATISTICS'
INSERT INTO DDL_ChangeEvents
(
EventType,
ObjectName,
ObjectType,
[sql] ,
Session_IPAddress
)
SELECT @eventdata.value('(/EVENT_INSTANCE/EventType)[1]',
'nvarchar(200)'),
@eventdata.value('(/EVENT_INSTANCE/ObjectName)[1]',
'nvarchar(200)'),
@eventdata.value('(/EVENT_INSTANCE/ObjectType)[1]',
'nvarchar(200)'),
@eventdata.value('(/EVENT_INSTANCE/TSQLCommand)[1]',
'nvarchar(max)'), client_net_address
FROM sys.dm_exec_connections WHERE session_id=@@SPID
;

Well I won't bother to record CREATE STATISTIC events hence there is an IF block to skip this event. I get the IP Address from sys.dm_exec_connections DMV which has client_net_address column.
Now create/drop/alter table (also via SSMS) for example and query the DDL_ChangeEvents table to see what happened.

There is a clever way of rebuild indexes

2008-10-05T04:44:00.000-07:00

Hi everybody.
It seems like I am seeing more and more inquires from our clients asking for help solving performance related issues with rebuilding indexes. All of them (or almost all of them) have been using Maintanace Plan Rebuild/Reorganize Index Task. We have lots of clients who have pretty big databases(>200GB) and have not hired yeat a DBA:-).They used to use this task and specify all tables as well as all databases, moreover, one client used to run such tasks in the middle of work day. It leads to locks on tables and performance decreasing. I would also notice you to not cancelling the task as SQL Server will rolback the whole transactions and you are about to wait a lot of time. Just let the task to complete. I suggested instead of running the task, first, identify fragmented indexes on tables that have more than 1000 pages.

DECLARE @RebuildStatement nvarchar(4000)
DECLARE RebuildStatements CURSOR LOCAL FAST_FORWARD
FOR
SELECT 'ALTER INDEX '+i.name+ ' ON '+
OBJECT_NAME(i.object_id)+' REORGANIZE;'
FROM
sys.dm_db_index_physical_stats(db_id(), NULL, NULL, NULL, 'DETAILED') phystat inner JOIN sys.indexes i
ON i.object_id = phystat.object_id
AND i.index_id = phystat.index_id WHERE phystat.avg_fragmentation_in_percent > 40
and page_count>=1000

OPEN RebuildStatements
WHILE 1 = 1
BEGIN
FETCH NEXT FROM RebuildStatements INTO @RebuildStatement
IF @@FETCH_STATUS <> 0 BREAK
EXEC(@RebuildStatement)
END
CLOSE RebuildStatements
DEALLOCATE RebuildStatements

The above SELECT generates a simple script to REORGANIZE (change to REBUILD) indexes and EXECUTES the dynamic sql. As you probaly know this script has to be run on SQL Server 2005/2008 and do not forget about really great feature such rebuilding indexes ONLINE. For more details please see BOL.

To speak kindly about RedGate

2008-09-15T00:17:00.000-07:00

All of us are aware that restore database with different collation may cause a headache. That what happened to one of our databases where one developer created a database without to pay attention about what kind of data he is going to deal with. Ok,we can set COLLATION even per column , but what if you have lots of tables to be altered with new COLLATION. Well, we can use either import/export or SSIS package or perhaps write some T-SQL script to do the job , however I'd like to tell you how easy to get a new database with desired COLLATION by using RedGate tool.
1) CREATE DATABASE dbname COLLATE 'your desired collation'
2) Open RedGate (SQL Compare)tool to move the structure of source db to the destination db. That's all.

I takes a few minutes even we had 25GB database. I'd strongly recommend to have a look at this great tool.(www.red-gate.com)

Computed column is PERSISTED?

2008-09-01T23:32:00.000-07:00

Hi everybody.
I'd like to share with you how important to define a computed column to be PESRISTED.
As you know from the BOL
/*
For columns specified as PERSISTED, the SQL Server 2005 Database Engine physically stores the computed values in the table and updates the values when any other columns on which the computed column depends are updated. By marking a computed column as PERSISTED, you can create indexes on computed columns defined on expressions that are deterministic, but not precise.
*/
I visited our client two days ago who has been experienced with performance issue for one of their very important query.One big table containes a computed column which SELECT statement is using for to return to the client. We have seen very high number of logical reads and TWO computer scalar iterators. For an obvious reason we define the computed column as PESRISTED and performance was increased dramatically.Moreover, create an index on computed column and see how perfromance will be increased more..

See demo script to see how it is affected.

CREATE TABLE t(c INT NOT NULL identity(1,1) PRIMARY KEY,
c1 AS '00'+cast(c AS VARCHAR(100)))

SET NOCOUNT ON
INSERT INTO t DEFAULT VALUES
GO 100000

SET STATISTICS IO ON
SET STATISTICS PROFILE ON
SELECT c1 FROM t
--Table 't'. Scan count 1, logical reads 1250, physical reads 0, read-ahead reads 8, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

--select c1 from t
|--Compute Scalar(DEFINE:([master].[dbo].[t].[c1]=[master].[dbo].[t].[c1]))
|--Compute Scalar(DEFINE:([master].[dbo].[t].[c1]='00'+CONVERT(varchar(100),[master].[dbo].[t].[c],0)))
|--Clustered Index Scan(OBJECT:([master].[dbo].[t].[PK__t__1446FBA6]))

SET STATISTICS IO ON
SET STATISTICS PROFILE ON

SELECT c1 FROM t
--Table 't'. Scan count 0, logical reads 50, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
--select c1 from t
|--Compute Scalar(DEFINE:([master].[dbo].[t].[c1]=[master].[dbo].[t].[c1]))
|--Clustered Index Scan(OBJECT:([master].[dbo].[t].[PK__t__1446FBA6]))

DROP TABLE t

String or binary data would be truncated

2008-08-19T04:49:00.000-07:00

Hi folks.
I named this article with this pretty famous error message. I'm sure that everybody has seen this error at least once.This week I visited our client who asked me very intresting question. They have a table with many columns that most of them defined as VARCHAR(n) datatype. One of the query has been failing with below error.

"Msg 8152, Level 16, State 14, Line 1
String or binary data would be truncated.
The statement has been terminated."

Well ,we knew for sure that the user supplied a string which does not match with column datatype but real question is WHICH of so many columns? As I said, they have more than 50 columns and it was pretty difficult to identify on which column is failed.
My point is that it would be nice to have more information from this error message about which column is failing and I hope that MS will do something for the matter.

Cannot delete a job which is related to MP?

2008-07-28T00:42:00.000-07:00

Hi folks
If you are using SQL Server 2005 (SP) and used to build Maintanace Plans especially with adding subplans so you are probable seen the folowing error message
/*
Drop failed for Job ‘jobname’. (Microsoft.SqlServer.Smo)

The DELETE statement conflicted with the REFERENCE constraint “FK_subplan_job_id”. The conflict occurred in database “msdb”, table “dbo.sysmaintplan_subplans”, column ‘job_id’.
The statement has been terminated. (Microsoft SQL Server, Error: 547)
*/

So if you create a MP,SQL Server will create a job and SSIS which is refernced to the subplan as well as inserts the data into system tables in msdb database. (sysmaintplan_subplans,sysjobs_view,sysjobschedules).

Intresting is that if you execute a job it makes more insert into log table called sysmaintplan_log. All of these tables are linked through FK and PK relationships. The problem is when you try to delete a job it gives a Foreign Key errors until you manually remove those entries by the SQL tables.

Please see the link written by Jonas Kempas http://gudenas.com/2007/04/20/sql-server-2005-delete-maintenance-plan-error/ explains step by step how to delete not associated jobs.

I tried it and it worked just fine.

How do we open a large table?

2008-07-02T04:35:00.000-07:00

This question raised when we worked at the client's side by one of developers. They use SQL Server 2005 (SP2) and remembered that in SQL Server 2000 we can open the table via EM and then specify (All rows , Top rows...) do you remember?

I rarely use SSMS to open/edit tables data, and it seems that MS just removed this option. In SQL Server 2005 we have TABLESAMPLE clause that used to open a table with lots of rows. So we only needed to see what kind of data this psecific table has without open entire table or using TOP clause.

This example returns an approximate percentage of rows and generates a random value for each physical 8-KB page in the table.

SELECT * FROM table
TABLESAMPLE system(5 PERCENT)

I'd really advise you to read BOL about this great feature
ms-help://MS.SQLCC.v9/MS.SQLSVR.v9.en/udb9/html/8868e8fd-6c42-4171-9eab-a0e38cb1bfd3.htm

How do you maintain SharePoint databases?

2008-06-19T05:42:00.000-07:00

My latest post is about how maintaining MSDB database, today I would like to share with you some thoughts about maintaining Sharepoint databases.

Last week I visited our client who has been working with SharePoint (MOSS7) and started complain about performance of SQL Server. I identified long running queries as well as very good number of deadlocks that happened every hour to databases which belong to SharePoint product. However , when I recommend to add some indexes on the tables people start almost crying not to do that as they were told that SharePoint databases are self managed product and DBA should not be touched it at all. It looks to me very strange , but that was my first experience with MOSS and I decided to do some searching on internet.

I found a couple of documents (even published by MS) to read them as if Sharepoint SQL Server performance can be managed by Shrinking & Defragging the DB.
I also asked some Sharepoint people and they say that accessing the database directly, changing anything on their databases
aside from what's provided out of the box, etc. is not supported unless you
do it thru the Sharepoint API. Hmm.... looks strange , does not?

Finally I ended up with sample script to identify very fragmented indexes and running ALTER INDEX index_name ON tablename REORGANIZE;

PS.
I could not imagine a customer being unwilling to create whatever indexes are
necessary to ensure reasonable performance of a production Sharepoint
system.

How do you maintain MSDB database?

2008-05-18T04:17:00.000-07:00

Last week I visited our client who has pretty big databases and performs BACKUP LOG ..operation on almost all user databases. Now, one of the most critical databases got corrupted and the DBA was pretty confident that he won't loose any data (as he had backup of log file) and brings the database from the backup within 10-12 minutes.

They also have very well written stored procedure that does RESTORE DATABASE based on name of the database and number of log files to be restored. They run the stored procedure and it has been running for almost 5 hours till DBA canceled the process. What happened? Why it has taken so much time? I thought about it and asked him a question,:-"Have you ever cleared backup history?", he replied that he hasn't. Then we checked backupset system database that contained more than one million rows!!!!

I remember SQL Server MVP Geoff N.Hiten wrote the blog about the issue and I even posted a comment on.Please check the following article
http://weblogs.sqlteam.com/geoffh/archive/2008/01/21/MSDB-Performance-Tuning.aspx

It tooks only 3 minutes to run a script that create indexes and about 15 munutes to run
use msdb

go

declare @OldestDate datetime

set @OldestDate = getdate() -100

exec sp_delete_backuphistory @OldestDate

Now that it is finished , our RESTORE command took only 12 minutes to complete.
I'd like to point out how important is to clear backup history (Fortunately, in SQL Server 2005 we have builtin taks to do the job) as on time 'X' you will succefully restore a needed database.

Getting row count of table in SQL Server 2005

2008-05-05T00:44:00.000-07:00

It is much easier in SQL Server 2005 to get row count per table.

SELECT
[TableName]=tbl.name,
[RowCount] = SUM
(
CASE
WHEN (pt.index_id < 2) AND (au.type = 1) THEN pt.rows
ELSE 0
END
)
FROM
sys.tables tbl
INNER JOIN sys.partitions pt
ON tbl.object_id = pt.object_id
INNER JOIN sys.allocation_units au
ON pt.partition_id = au.container_id
GROUP BY
tbl.name ORDER BY [RowCount]DESC;

Getting next value

2008-05-04T02:05:00.000-07:00

Hi folks. I'd like to share with you some technique to get a next value from the table means to create your own sequence mechanism.
As we all know that an Identity property may have gaps, so use the below script to retrieve a next value. We create a table with a one row and a only one column which holding the last used sequence value.

CREATE TABLE seq(col int not null);
INSERT INTO seq values(0);
go
SELECT * FROM seq

CREATE PROC spget_nextseq @next_val AS INT OUTPUT
AS
UPDATE seq SET @next_val= col = col + 1;
go

-- usage
DECLARE @i as int;
EXEC spget_nextseq @i output;
SELECT @i;

Note, in OLTP application where many connections run this script you may end up with deadlocks. One way to pevent it is using lock hint called TABLOCK

ALTER PROC spget_nextseq @next_val AS INT OUTPUT
AS
UPDATE seq SET @next_val= col = col + 1 FROM seq WITH (TABLOCK);
go

Quick look at IN predicate

2008-04-20T22:54:00.000-07:00

There are lots of articles and techniques onnthe internet about how to deal with delimited parameters. Last Wed I was visited a client that asked to write a quick query to return the data based on delimited values . Here we go

Use pubs

DECLARE @t VARCHAR(50)
SET @t = ('Bennet,smith')

SELECT *
FROM authors
WHERE ',' + @t + ',' LIKE '%,' + au_lname + ',%'

SELECT * FROM
authors WHERE CHARINDEX(',' + au_lname + ',',','+ @t+',')>0

As you can imagine, performance will be horrible.

SQL Server Agent Jobs duration Report

2008-03-17T03:51:00.001-07:00

Hi folks. I would like to share with you the following simple script to show us jobs duration report in SQL Server 2005. I manipulated with INTERGER values stored by SQL Server to convert them into DATETIME/CHAR(8) datatypes to represent the data.
Thanks to SQL Server MVP Peter Ward provided me with StartTime calculation.

WITH job_duration_view
AS
(
SELECT name,
StartTime = CONVERT(DATETIME, RTRIM(last_run_date)) +
(last_run_time * 9 + last_run_time % 10000 * 6 + last_run_time % 100 * 10 + 25 * last_run_duration) / 216e4 ,
CONVERT(CHAR(8),DATEADD(ss,last_run_duration,CAST(last_run_date AS CHAR(8))),114)
AS duration
FROM msdb.dbo.sysjobservers js
JOIN msdb.dbo.sysjobs j ON j.job_id = js.job_id
WHERE last_run_date >0 AND last_run_time >0
) SELECT name AS job_name,StartTime,
StartTime -'19000101'+Duration AS EndDate ,Duration
FROM job_duration_view

Change collation in tempdb

2008-03-09T08:10:00.000-07:00

What's happening if you installed SQL Server instance with a collation that is different from a database collation? We just started testing our production application and everything seem to work well, however one of our stored procedure inserts hebrew characters into temporary table and then after some operations the data get insertded into a real table. Guess what wee have seen in the database? Right,we have seen '????' symbols. Sure, if you do not use temporary table and insert the data directly into a permanent table you will see the right characters. Someone said that we should decline of using temporary table and insert the data into 'temporary' permanent table. Another guy said that we should run ALTER DATABASE tempdb command to change COLLATION, but as we know you cannot run this statement on system databases.
The error is
Msg 3708, Level 16, State 5, Line 1
Cannot alter the database 'tempdb' because it is a system database.

So what is the solution? Well , use the REBUILDDATABASE option in Setup.exe or re-install the instance. Fortunately, the whole story happened on developing machine and we did not forget to install PRODUCTION server with right collation:-).

Just wanted to note you how important is to choose the 'right' collation while installing production server.

Scheduling database backup in SQL 2005 Express Edition

2008-03-02T01:51:00.000-08:00

I have been recently visited client who uses SQL Server 2005 Express Edition for his business. As you know that edition is supplied without SQL Server Agent (which was in MSDE). So,the question is how to schedule backup of the database in that case? I would like to thank to Jasper Smith, who wrote these greate articles

Automating Database maintenance in SQL 2005 Express Edition Part I
http://www.sqldbatips.com/showarticle.asp?ID=27

Automating Database maintenance in SQL 2005 Express Edition Part II
http://www.sqldbatips.com/showarticle.asp?ID=29

CREATE PROCEDURE usp_BackupDatabases
@dbname SYSNAME,
@backup_path VARCHAR(100)
AS
DECLARE @dt VARCHAR(20) ---date of db backup
DECLARE @FileName VARCHAR(50) ---build path and dbname
DECLARE @sql VARCHAR(100) ---build final sql statement

SET @dt =CONVERT(VARCHAR(15),GETDATE(),112)
SELECT @FileName = @backup_path+@dbname + @dt+'.bak'

SET @sql = N'BACKUP DATABASE [' + @dbname + '] TO DISK = ''' + @FileName + ''''
EXEC(@sql)

The above is simple stored procedure that accepts two parameters and performs BACKUP of the database.Now that, we have created stored procedure lets put the call of the sp into a text file ( I use userdbfullbackup.sql) and place the file on filesystem.
usp_BackupDatabases @dbname='$(DB)',@backup_path='$(BACKUPFOLDER)'

SQLCMD utility 'knows' to accept parameters (-v),so we can easily provide parameters.

--Usage
EXEC master..xp_cmdshell 'SQLCMD -S URID\SQLSERVERDEV2005 -i"c:\userdbfullbackup.sql" -v DB="Northwind" -v BACKUPFOLDER="c:\"'

But wait , till now we just written a stored procedure that backups our databases and tested it , so what is about scheduling?

Windows Servers/XP have Scheduled Tasks Wizard
You can refernce to below KB
http://support.microsoft.com/default.aspx?scid=kb;en-us;308569&sd=tech

Please see Jasper's explanation ( do not want repetition).
http://www.sqldbatips.com/showarticle.asp?ID=27

ISNULL changes to NOT to allow NULLs

2008-02-24T04:55:00.000-08:00

Hi folks. Actually, it is pretty old trick/feature I would like to show you. Let's say we have a table called 'Test1' with one column as VARCHAR(10) which allows inserting NULLs. Well,we want to change this to NOT allow NULLs and I'm also going to change datatype for this column at the same time ,so just for testing let's assume that 'Test1' does not contain alphabetic values.

CREATE TABLE Test1
(
Col1 varchar(10) NOT NULL,
)
INSERT INTO Test1 VALUES('1')

EXEC sp_rename 'Test1', 'Test1_Old'
GO
SELECT ISNULL(CAST(Col1 AS int), 0) AS Col1 ---Makes Col1 not to allow NULLs
INTO Test1
FROM Test1_Old

DROP TABLE Test1_Old

Make sure that you do not have duplicates, so you can create a constraint on the column.
/*
ALTER TABLE Test1
ADD CONSTRAINT PK_Test1
PRIMARY KEY (Col1)
*/
GO

Percent of completion for some operations in SS2005

2008-02-12T00:56:00.000-08:00

Hi folks
I have seen many and many businesses today have already upgraded to SQL Server 2005 and at this time I would like to share with you some of new features that this product gives us.
As you know SQL Server 2005 provides lots of dynamic management views (DMV) to get access to internal behaviour of processes or commands. If I remember well, someone of MVPs has already published about this feature , so anywhere..

Take a look at this script. Run for example BACKUP DATABASE comand and in another session run the below script. It provides you with start time column , percent of complete and what possible completion time is...(estimated_completion_time ) COOL, right?.

SELECT TOP 2 start_time,
percent_complete ,estimated_completion_time
FROM sys.dm_exec_requests
ORDER BY start_time desc

Check it out with those commands

DBCC CHECKDB
DBCC SHRINKDATABASE
DBCC SHRINKFILE
BACKUP DATABASE
ROLLBACK

Tables that do not have Clustered Index

2008-01-21T04:24:00.000-08:00

I'd like to share with you this script that returns name of tables that do not have Clustered Index defined on.You may ask, why do we need to create a Clustered Index on the table at all? Well,I think it is a subject for another blog , my experience is that every table should have clustered index.

--SQL Server 2005 (SP2)

SELECT name AS object_name FROM sys.objects WHERE name NOT IN
(
SELECT o.name AS object_name
FROM sys.indexes i
JOIN sys.objects o
ON i.object_id = o.object_id
WHERE objectproperty(o.object_id,'IsMSShipped') =0
AND I.type=1 ---Clustered Index
) AND objectproperty(sys.objects.object_id,'IsMSShipped') =0 ---Only user objects
AND objectproperty(sys.objects.object_id,'IsTable') =1 ---Only tables
ORDER BY name

Columns_Updated() within a trigger

2008-01-08T06:14:00.000-08:00

Happy New Year to everyone. Once someone asked me about what columns are affected by the UPDATE statement. As you probaly know that in SQL Server 2005 we can do that in more elegant way, but for SQL Server 2000 I decided perfrom some testing.

CREATE TABLE Test (c1 INT NOT NULL PRIMARY KEY, c2 CHAR(1),c3 INT)
GO
--Insert some data
INSERT INTO Test VALUES (1,'A',100)
INSERT INTO Test VALUES (2,'B',500)

--Create a trigger
CREATE TRIGGER tr_my_trigger ON Test FOR UPDATE
AS
DECLARE @ColumnID INT
DECLARE @Columns VARCHAR(8000)
SET @Columns = SPACE(0)
SET @ColumnID = 1
WHILE @ColumnID <= (SELECT COUNT(*)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = 'Test')
BEGIN
IF (SUBSTRING(COLUMNS_UPDATED(),(@ColumnID - 1) / 8 + 1, 1)) &
POWER(2, (@ColumnID - 1) % 8) =
POWER(2, (@ColumnID - 1) % 8)
SET @Columns = @Columns + CAST(@ColumnID AS VARCHAR) + ','
SET @ColumnID = @ColumnID + 1
END
PRINT 'Updated columns are :' + @Columns

---
UPDATE Test SET c2='D',c3=2000 WHERE c1=1
--Updated columns are :2,3
We see that the second and the third column were affected according to the INFORMATION_SCHEMA.COLUMNS ordered by ORDINAL_POSITION (check it out). It is not comfortable to see those numbers , but with a little effort you can get also name of columns from the INFORMATION_SCHEMA.COLUMNS.

I suggest you not to run/use that on the production server as it may hurt performance.

Rebuild index, when does it complete?

2007-12-17T22:22:00.000-08:00

I have recently worked on the client site (SQL Server 2000) to help on optimization. They have a very huge table (about 250 million rows) and perfom rebuild index every day at night. That jobs has beeen completed in two hours or so till this monday. The job was in 'executin status' for more than 10 hours and imagine what happened to the LOG file. My first thing was it would be really helpfull to have any system table that shows a percentage or whatever about the rebuild index process (something similar to what we have in SQL Server 2005 for some commands (BACKUP DATABASE) to see percent's complete )as I did not intend to KILL the process.Finally we got on and the process was completed. I suggested to the client rebuilding only heavy fragmented indexes especially if you have such huge table. Also , it is unnecessary to rebuild that table every day (it depends on your business requirements) ,so in their case , it is enough once a week.

In SQL Server 2005 for below operations you are able to see completion time or percent of completion.

DBCC CHECKDB
DBCC SHRINKDATABASE
DBCC SHRINKFILE
BACKUP DATABASE
ROLLBACK

SELECT TOP 2 start_time,
percent_complete ,estimated_completion_time
FROM sys.dm_exec_requests
ORDER by start_time desc

Shrink TempDB database

2007-12-02T04:51:00.000-08:00

I have been recently worked for some client who has huge tempdb size (SQL Server 2000). As you probably know, you cannot reduce physical size of tempdb without restart MSSQLServices. Using the command DBCC SHRINKFILE to shrink the individual tempdb files you must run DBCC SHRINKFILE command while no other activity occurs in the tempdb database. To make sure that other processes cannot use tempdb while DBCC SHRINKFILE executes, you must restart SQL Server in the single user
mode.
So I would like to show some stepS we ended up. The client completely understands the problem cause tempdb database to be growing and I hope it is not short time solution.

1) Stop SQL Server. Open a command prompt, and then start SQL Server by
typing the following command:

"sqlservr -c -f" (remove the quotation)

The -c and -f parameters force SQL Server to start in a minimum
configuration mode with a tempdb size of 1 MB for the data file and 0.5 MB
for the log file.
Connect to SQL Server with Query Analyzer, and then issue the following
TSQL commands for the tempdb database files that you need to shrink

use tempdb
go

dbcc shrinkfile (tempdev, 'target size in MB') --for the primary data file
or
dbcc shrinkfile (templog, 'target size in MB')--for the log file

As MS says
"An advantage of DBCC SHRINKFILE is that it can reduce the size of a file to
a size smaller than its original size. A limitation of DBCC SHRINKFILE is that you cannot make the database smaller than the size of the model database."

Non-updateable views

2007-11-13T05:41:00.000-08:00

As we know from BOL
"Updatable views can modify more than one table involved in the view. The DELETE, INSERT, and UPDATE statements can reference a view as long as SQL Server can translate the user's update request unambiguously to updates in the base tables referenced in the view's definition"

But I would like to share with you one method to prevent updating view that based on single table. Let's see the following

CREATE TABLE1 (col1 int,col2 int,col3 int,col4 int,col5 int)
GO
INSERT INTO t1 VALUES (1,1,11,2,12)
INSERT INTO t1 VALUES (8,10,15,25,55)
INSERT INTO t1 VALUES (9,1,11,2,5,81)
GO
CREATE VIEW V1 WITH VIEW_METADATA
AS
SELECT
col1,
col2+0 AS col2,
col3,
col4+0 AS col4,
col5
FROM T1
GO

--Usage

UPDATE v1 SET col1=100 where col2=10
The above statement is succeed, but what happened to the next one

UPDATE v1 SET col2=1000 where col3=15
--Update or insert of view or function 'v1' failed because it contains a derived or --constant field.

You cannot derived column ,as it should be referring to the update, not the view.
In SQL Server 2000, you can handle it using an INSTEAD OF trigger.

Reading Registry

2007-10-11T06:08:00.000-07:00

Sometimes we need to read registry by using T-SQL . I would like to show you some scripts to do the job.

--Getting MDAC
EXEC master..xp_regread
N'HKEY_LOCAL_MACHINE',
N'Software\Microsoft\DataAccess',
N'Version'

--Getting an account SQL Server runs under
DECLARE @serviceaccount varchar(100)
EXECUTE master.dbo.xp_instance_regread
N'HKEY_LOCAL_MACHINE',
N'SYSTEM\CurrentControlSet\Services\MSSQLSERVER',
N'ObjectName',
@ServiceAccount OUTPUT,
N'no_output'

SELECT @Serviceaccount

---Getting SQL Server path's installation and the data path
DECLARE @DataPath nvarchar( 512 ) , @SQLPath nvarchar( 512 ) ,
@ToolsPath nvarchar( 512 )
EXECUTE sp_MSget_setup_paths @SQLPath OUT , @DataPath OUT;
PRINT 'SQLPath : ' + QUOTENAME( @SQLPath , '"');
PRINT 'DataPath : ' + QUOTENAME( @DataPath , '"');

How many times stored procedures have recompiled?

2007-10-08T07:23:00.000-07:00

I have recently visited our client and we found out that He has a very high number of stored procedures that have recompiled. Certainly, we have observed some perfomance decreasing. Fortunately , the client has alredy upgraded to SQL Server 2005(SP2) and apart from running SQL Server Profile to indentify those stored procedures we came out with the following query.

It gives you the top 10 stored procedures that have been recompiled.

select *
from sys.dm_exec_query_optimizer_info

select top 10
sql_text.text,
sql_handle,
plan_generation_num,
execution_count,
dbid,
objectid
from
sys.dm_exec_query_stats
cross apply sys.dm_exec_sql_text(sql_handle) as sql_text
where
plan_generation_num >1
order by plan_generation_num desc

Note that the plan_generation_num indicates the number of times the stored procedure has recompiled. Very cool feature.

RESTORE LOG file and WITH FILE behaviour

2007-10-01T01:14:00.000-07:00

Recently I have been visited one of the our clients and have been asked the following question.
"What if I have 3 fileid within a backup file and when restore the log being done, I do not specify WITH FILE option,so do I get latest Log's data?". To answer the question I decide to conduct some test script.

CREATE DATABASE demo
GO
ALTER DATABASE demo SET RECOVERY FULL
GO
CREATE TABLE demo..t1(c1 INT NOT NULL PRIMARY KEY)
--insert one row
insert demo..t1 VALUES (1)
--BACKUP the data
BACKUP DATABASE demo TO DISK = 'c:\temp\demo.bak' WITH INIT
GO
insert demo..t1 VALUES (2)
--start BACKUP log data
BACKUP log demo TO DISK = 'c:\temp\demo_log.bak' WITH INIT
GO
insert demo..t1 VALUES (3)
--- BACKUP log FILE at this time WITH FILE =2
BACKUP log demo TO DISK = 'c:\temp\demo_log.bak' WITH NOINIT
GO
insert demo..t1 VALUES (4)
--- BACKUP log FILE at this time WITH FILE =3
BACKUP log demo TO DISK = 'c:\temp\demo_log.bak' WITH NOINIT

--Lets do RESTORE and see what is going on
RESTORE DATABASE demo FROM DISK = 'c:\temp\demo.bak' WITH FILE = 1, NORECOVERY
RESTORE LOG demo FROM DISK = 'c:\temp\demo_log.bak' WITH FILE = 1, NORECOVERY
RESTORE LOG demo FROM DISK = 'c:\temp\demo_log.bak' WITH RECOVERY ----Do you expect the last one?

GO
SELECT * FROM demo..t1
c1
-----------
1
2
It did not return rows 3 and 4.

So that means if you DO NOT specify WITH FILE in RESTORE LOG file you DO NOT get latest FILEID but only the first one.

Insert data into a Text File

2007-09-23T00:51:00.000-07:00

As you propably know that MS introduced in SQL Server 2005 a new command utility called SQLCMD. In this article I would like to show the difference between SQLCMD and "old fashioned" OSQL in terms of formating the text file's output.

If you run these two statements and look at the files , you'll see that SQL Server inserts three dotted lines in myoutput2.txt, not so good for reading. Opposite, in myoutput1.txt you see very good reading format.

EXEC master..xp_cmdshell 'SQLCMD -S URID\SQLSERVERDEV2005 -E -Q "SELECT TOP 10 * FROM pubs.dbo.authors" -b -o c:\myoutput1.txt', no_output
EXEC master..xp_cmdshell 'OSQL -S URID\SQLSERVERDEV2005 -E -Q "SELECT TOP 10 * FROM pubs.dbo.authors" -b -o c:\myoutput2.txt', no_output

Despite that SQL Server 2005 supports OSQL utility, I strongly recommend you using SQLCMD.

Disable all Foreign Keys

2007-09-16T22:11:00.000-07:00

Recently I was told to clear up all user tables in the client's database SQL Server 2000 SP3.There are lots of contsrtaints and for this purpose we would like to disable them in order to delete all data. Try it on Northwind database before using on production. If you want to enable constraints ,simple replace NOCHECK with CHECK.

DECLARE @TableName nvarchar(257),
@ForeignKeyConstraintName sysname,
@SQLStatement nvarchar(4000)

DECLARE TableList CURSOR
LOCAL FAST_FORWARD READ_ONLY FOR
SELECT
QUOTENAME(TABLE_SCHEMA) +
N'.' +
QUOTENAME(TABLE_NAME) AS TableName,
QUOTENAME(CONSTRAINT_NAME) AS ForeignKeyConstraintName
FROM
INFORMATION_SCHEMA.TABLE_CONSTRAINTS
WHERE
CONSTRAINT_TYPE = 'FOREIGN KEY' AND
OBJECTPROPERTY
(
OBJECT_ID
(
QUOTENAME(TABLE_SCHEMA) +
N'.' +
QUOTENAME(TABLE_NAME)
),
'IsMSShipped') = 0

OPEN TableList
WHILE 1 = 1
BEGIN
FETCH NEXT FROM TableList INTO
@TableName,
@ForeignKeyConstraintName
IF @@FETCH_STATUS = -1 BREAK
SET @SQLStatement =
N'ALTER TABLE ' +
@TableName +
N' NOCHECK CONSTRAINT ' +
@ForeignKeyConstraintName
RAISERROR (@SQLStatement, 0, 1) WITH NOWAIT
EXEC sp_executesql @SQLStatement
END
CLOSE TableList
DEALLOCATE TableList
GO

Getting an index's information

2007-08-28T02:42:00.000-07:00

I was asked by one of the client to get all info about the indexes i.e name of the index , column's name, table's name etc. Please take a look at following script to return the data for SQL Server 2000.
use pubs

SELECT tbl = object_name(i.id), i.name as index_name, c.name as column_name,
isunique = indexproperty(i.id, i.name, 'IsUnique'),
isclustered = indexproperty(i.id, i.name, 'IsClustered'),
constrtype = CASE o.type
WHEN 'PK' THEN 'PRIMARY KEY'
WHEN 'UQ' THEN 'UNIQUE'
END
FROM sysindexes i
JOIN syscolumns c on i.id = c.id
JOIN sysindexkeys k on i.id = k.id
and i.indid = k.indid
and c.colid = k.colid
LEFT JOIN sysobjects o ON o.name = i.name
AND o.xtype in ('PK', 'UQ')
AND o.parent_obj = i.id
WHERE indexproperty(i.id, i.name, 'IsHypothetical') = 0
AND indexproperty(i.id, i.name, 'IsStatistics') = 0
AND indexproperty(i.id, i.name, 'IsAutoStatistics') = 0
AND objectproperty(i.id,'IsMSShipped')=0
ORDER BY tbl, i.name, k.keyno

For SQL Server 2005 I use the script written by Kalen Delaney.Remember there is new type of index called INCLUDE.

CREATE VIEW get_index_columns
AS
SELECT object_name(ic.object_id) as object_name , index_name = i.name,
'column' = c.name,
'column usage' = CASE ic.is_included_column
WHEN 0 then 'KEY'
ELSE 'INCLUDED'
END
FROM sys.index_columns ic JOIN sys.columns c
ON ic.object_id = c.object_id
AND ic.column_id = c.column_id
JOIN sys.indexes i
ON i.object_id = ic.object_id
AND i.index_id = ic.index_id

After creating the view, you can select from it, and it will give you the
KEY columns and the INCLUDED columns in all the indexes in all the tables.
Or, you can add a WHERE clause for your own table or index:

SELECT * FROM get_index_columns
WHERE object_name = 'mytable'

Foreign Key dependency

2007-08-14T01:20:00.000-07:00

Some time ago I was asked to truncate lots of tables in client's database.As such,all those tables have FK and PK dependency. You get the error like that if you run TRUNCATE TABLE Categories in Nortwind database as an example.
--Cannot truncate table 'Categories' because it is being referenced by a FOREIGN KEY constraint.

So , there is no such great problem as you know the database structure and can easily indetify parent-child tables. But what if you don't?
I'd like to show the function (sadly, I don't remember by whom it is written) that returns the tables ordered by its level. Level 0 means that it is refernced table and should be truncated last.

CREATE VIEW VFKs
AS

SELECT
FK.TABLE_SCHEMA AS child_table_schema,
FK.TABLE_NAME AS child_table_name,
PK.TABLE_SCHEMA AS parent_table_schema,
PK.TABLE_NAME AS parent_table_name
FROM INFORMATION_SCHEMA.REFERENTIAL_CONSTRAINTS AS RC
JOIN INFORMATION_SCHEMA.TABLE_CONSTRAINTS FK
ON RC.CONSTRAINT_SCHEMA = FK.CONSTRAINT_SCHEMA
AND RC.CONSTRAINT_NAME = FK.CONSTRAINT_NAME
JOIN INFORMATION_SCHEMA.TABLE_CONSTRAINTS PK
ON RC.UNIQUE_CONSTRAINT_SCHEMA = PK.CONSTRAINT_SCHEMA
AND RC.UNIQUE_CONSTRAINT_NAME = PK.CONSTRAINT_NAME
GO

CREATE FUNCTION dbo.fn_get_tree_of_tables()
RETURNS @tree TABLE
(
table_schema SYSNAME NOT NULL,
table_name SYSNAME NOT NULL,
lvl INT NOT NULL,
PRIMARY KEY(table_schema, table_name)
)
AS
BEGIN

DECLARE @lvl AS INT
SET @lvl = 0

-- top level tables
INSERT INTO @tree
SELECT DISTINCT
parent_table_schema,
parent_table_name,
@lvl
FROM VFKs AS P
WHERE NOT EXISTS
(SELECT *
FROM VFKs AS C
WHERE C.child_table_schema = P.parent_table_schema
AND C.child_table_name = P.parent_table_name)

WHILE @@rowcount > 0
BEGIN

SET @lvl = @lvl + 1

-- non top level tables
INSERT INTO @tree
SELECT DISTINCT
child_table_schema,
child_table_name,
@lvl
FROM VFKs AS C
JOIN @tree AS P
ON P.table_schema = C.parent_table_schema
AND P.table_name = C.parent_table_name
AND lvl = @lvl - 1
WHERE NOT EXISTS
(SELECT *
FROM @tree AS T
WHERE T.table_schema = C.child_table_schema
AND T.table_name = C.child_table_name)

END

SET @lvl = 0

-- tables with no fks
INSERT INTO @tree
SELECT TABLE_SCHEMA, TABLE_NAME, @lvl
FROM INFORMATION_SCHEMA.TABLES AS TB
WHERE NOT EXISTS(
SELECT *
FROM @tree AS TR
WHERE TB.TABLE_SCHEMA = TR.table_schema
AND TB.TABLE_NAME = TR.table_name)
AND
TB.TABLE_TYPE = 'BASE TABLE'

RETURN

END
GO

Here's the results of an invocation of the function in Northwind:

SELECT * FROM dbo.fn_get_tree_of_tables() AS T
ORDER BY lvl

table_schema table_name lvl
------------ ----------------------------------- ---
dbo Categories 0
dbo CustomerDemographics 0
dbo Customers 0
dbo Employees 0
dbo Region 0
dbo Shippers 0
dbo Suppliers 0
dbo Products 1
dbo Territories 1
dbo Orders 1
dbo CustomerCustomerDemo 1
dbo Order Details 2
dbo EmployeeTerritories 2

Object's modified date

2007-08-06T05:30:00.000-07:00

There are lots of businesses have already upgraded to SQL Server 2005.At this time I'd like to take a look at great feature that we have in SQL Server 2005. I have been asked so many times how to know WHEN does the object change? It is a headache in SQL Server 2000. Fortunately, now we can answer that question.

Let's say the user changed number of stored procedures.Using the following statement we can easily get the data.

SELECT name,modify_date FROM sys.procedures
ORDER BY modify_date desc

But let's take it a little bit futher. I'd like to know when does user update the table last time?

CREATE TABLE dbo.uri(c INT NOT NULL PRIMARY KEY)
GO
INSERT INTOdbo.uri VALUES(10)
--we got one entry

SELECT object_name(object_id) AS name ,last_user_update
FROM sys.dm_db_index_usage_stats
WHERE database_id = db_id( 'demo' )

---try to change the data
UPDATE dbo.uri SET c =50 WHERE c =10

--you'll see that last_user_update is changed

SELECT object_name(object_id) AS name ,last_user_update
FROM sys.dm_db_index_usage_stats
WHERE database_id = db_id( 'demo' )

So ,you can play with these Dynamic Management Views to obtain more info.

Columns List of Views

2007-07-10T00:23:00.000-07:00

One of myclient asked me nice question about getting columns list for views. Well, please take a look at following solution I came up with.

The first view just holds the data about view and their columns and as you can see it is not easy to read it at all.

CREATE VIEW v_getcolumns
AS
SELECT INFORMATION_SCHEMA.TABLES.TABLE_NAME ,COLUMN_NAME
FROM INFORMATION_SCHEMA.TABLES JOIN INFORMATION_SCHEMA.COLUMNS
ON INFORMATION_SCHEMA.TABLES.TABLE_NAME=INFORMATION_SCHEMA.COLUMNS.TABLE_NAME
WHERE TABLE_TYPE = 'VIEW' AND
OBJECTPROPERTY(
OBJECT_ID(
QUOTENAME(INFORMATION_SCHEMA.TABLES.TABLE_SCHEMA) + '.' +
QUOTENAME(INFORMATION_SCHEMA.TABLES.TABLE_NAME)
), 'IsMSShipped') = 0

The final step I build a crosstab to contacanate the names of the columns per view based on count.I ened upi with 10 columns per view as maximum number, you can modify it for your needs.The cool thing is if you specify '*' in SELECT statement within a view and running the following query you get all columns of the table.

SELECT TABLE_NAME ,MAX(CASE WHEN rnk = 1 then COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 2 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 3 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 4 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 5 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 6 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 7 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 8 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 9 THEN ', ' + COLUMN_NAME ELSE '' end)
+MAX(CASE WHEN rnk = 10 THEN ', ' + COLUMN_NAME ELSE '' end) AS columns
FROM
(
SELECT A.TABLE_NAME ,A.COLUMN_NAME,COUNT(*)rnk
FROM v_getcolumns A, v_getcolumns B
WHERE A.TABLE_NAME=B.TABLE_NAME AND
A.COLUMN_NAME>=B.COLUMN_NAME
GROUP BY a.TABLE_NAME ,a.COLUMN_NAME
) AS der
GROUP BY TABLE_NAME

T-SQL and CLR types do not match

2007-07-08T05:09:00.000-07:00

As you know in SQL Server 2005 you can reference to .NET classes doing some operations that in T-SQL is realy hard to perform. For example if you need to format very long and 'complicated' string value , so by using .NET String method you can easily achive it. But I'd like to show you another example actually posted by Itzik Ben-Gan on his website ,so I just modified his C# code to VB.NET that is invoking the ToString method of the SqlDateTime datatype with a format string.

Imports Microsoft.SqlServer.Server
Imports System.Data.SqlClient
Public Class Valid
"<"SqlFunction(DataAccess:=DataAccessKind.Read, IsDeterministic:=True)">" _
Public Shared Function fn_format_datetime(ByVal dt As SqlTypes.SqlDateTime, _
ByVal formatstring As SqlTypes.SqlString) As String
If dt.IsNull Or formatstring.IsNull Then
Return formatstring.IsNull()
Else
Return dt.Value.ToString(formatstring.Value)
End If

End Function

End Class

I assume you are familiar with .NET technology , so I build a solution and try to reference to this function within my database.I created ASSEMBLY without problem but when I tried to create a function I got the error "CREATE FUNCTION for "fn_format_datetime1" failed because T-SQL and CLR types for return value do not match."

CREATE ASSEMBLY dateformat1 FROM 'C:\dateformat1 .dll'
with permission_set = SAFE
CREATE FUNCTION fn_format_datetime(@dt as datetime,
@formatstring as VARCHAR(100))RETURNS VARCHAR(100)
AS EXTERNAL NAME dateformat1.[dateformat1.classname].fn_format_datetime

As you see T-SQL datatype VARCHAR(n) does not match .NET STRING class. I should use NVARCHAR(n) datatype to get it working, did not know it.

CREATE ASSEMBLY dateformat1 FROM 'C:\dateformat1 .dll'
with permission_set = SAFE
CREATE FUNCTION fn_format_datetime(@dt as datetime,
@formatstring as NVARCHAR(4000))RETURNS NVARCHAR(4000)
AS EXTERNAL NAME dateformat1.[dateformat1.classname].fn_format_datetime

--Usage
SELECT
dbo.fn_format_datetime(GETDATE(), 'yyyyMMdd'),
dbo.fn_format_datetime(GETDATE(), 'yyyymmdd');

Trigger's definition

2007-06-26T02:33:00.000-07:00

I have had a clinet who wanted to script out the source of triggers on database level, so I'm talking about SQL Server 2005. Well as you probably know you can easily script the objects by using Tasks ---> Generate Scripts.. and moreover if you installed SP2 you will be able to script the object per file.

I would like to show you more simple way. You join two system tables on object_id column where sys.sql_modules table has a definition column which my client wanted to get.The WHERE condition restricts the ouptupt for triggers that were designed on database level.

SELECT definition FROM sys.sql_modules sm
INNER JOIN sys.triggers st ON
sm.object_id=st.object_id
WHERE parent_class_desc ='DATABASE'

Is Linked Server available?

2007-06-20T05:34:00.000-07:00

In SQL Server 2000 we have been checking an availability of linked server by using
xp_cmdshell along with 'ping' command.It worked fine but what does SQL Server 2005 propose to us?
There is a system stored procedure (sp_testlinkedserver )that accepts name of linked server as a parameter and returns the status of server.

Take a look at the script below.

DECLARE @srvr NVARCHAR(128), @statusval INT;
SET @srvr = 'Linked_Server';
BEGIN TRY
EXEC @statusval = sp_testlinkedserver @srvr;
END TRY
BEGIN CATCH
SET @statusval = SIGN(@@ERROR);
END CATCH;
IF @statusval <> 0
RAISERROR('Unable to connect to linked server. Try later!', 16, 2 );

Enabling AWE on Cluster environment

2007-06-03T04:38:00.000-07:00

I know many people ask this question. Recently I had a client wants me to enable AWE on the cluster which has two nodes (Active/Passive). He had only one instance of SQL Server 2000 (SP4) and 8 GB RAM and /3GB in the Boot.INI.

1)
Add /PAE parameter to Boot.INI file. PAE is the added ability of the IA32 processor to address more than 4 GB of physical memory.
2)
You need to install hotfix on SQL Server instance from Microsoft (http://support.microsoft.com/kb/899761) that fixs the bug "Not all memory is available when AWE is enabled".
3)
Make sure that domain account SQL Server is running under is added to "Lock pages in memory" local group policy.
4)
Finally we configured 6 GB for SQL Server as it was business requirements.

sp_configure 'show advanced options', 1
RECONFIGURE
GO
sp_configure 'awe enabled', 1
RECONFIGURE
GO
sp_configure 'max server memory', 6144
RECONFIGURE

5)Restart OS

Note: we need to run those steps on Active node where the SQL Server instance is running on. When it failed over but the binaries are loaded on all
nodes.

Linked servers and an Identity property

2007-05-16T02:46:00.000-07:00

Someone in sqlserver.programming group asked nice question about getting an IDENTITY from the linked server. I came up with the following solution. I assume that there is a table on the linked server that has one column defined as indentity property.

CREATE FUNCTION dbo.fn_get_remoteidentity()
RETURNS INT
AS
BEGIN
RETURN(
SELECT row_id
FROM OPENQUERY(
[put here your server name],
'SET NOCOUNT ON;
INSERT INTO dbname..usertable DEFAULT VALUES
ROLLBACK;
SELECT SCOPE_IDENTITY() AS row_id;') AS Der)
END
Another question was how to insert an INDENTITY property in the table on remote(linked) server?
Well in that case you will need to build dynamic SQL along with SET IDENTITY_INSERT
command.

DECLARE @sql NVARCHAR(400)
SELECT @sql = '
CREATE TABLE test(id INT IDENTITY(1,1))
SET IDENTITY_INSERT test ON
INSERT test DEFAULT VALUES
SET IDENTITY_INSERT test OFF
SELECT id FROM test'
--Usage
EXEC [LinkedServer].master.dbo.sp_executesql @sql

Global stored procedure

2007-05-13T02:46:00.000-07:00

I have my doubts that many of us have been used this system stored procedure(sp_MS_marksystemobject) to resolve the problems. But it exists and gives you an ability to return the data in the scope of selected database. Well, across number of databases having one table with the same structure and being in the scope of desired database you'll return the data (by using the following stored procedure) from the table located in desired database.To make the stored procedure to recognize the 'desired' data we use sp_MS_marksystemobject system stored procedure. BTW, it works as well as on SQL Server 2000 and on SQL Server 2005(SP2).

USE master
CREATE TABLE test (c1 VARCHAR(50))
INSERT test VALUES('master')
go
CREATE PROC sp_test AS
SELECT * FROM test
GO
USE northwind
CREATE TABLE test (c1 VARCHAR(50))
INSERT test VALUES('northwind')
USE pubs
CREATE TABLE test(c1 VARCHAR(50))
INSERT test VALUES('pubs')
USE pubs
EXEC sp_test --returns 'master'
USE master
EXEC sp_MS_marksystemobject sp_test
USE pubs
EXEC sp_test --returns 'pubs'
USE northwind
EXEC sp_test --returns 'northwind'

Find specific data type of the column

2007-05-09T00:42:00.000-07:00

Recently i was told to return tables which have FLOAT datatype defined on columns.The version of the product was SQL Server 2000 ,however it will work on SQL Server 2005 as well, so take a look at very very simple script.

SELECT columns.TABLE_NAME ,COLUMN_NAME
FROM information_schema.columns columns
JOIN information_schema.tables tables
ON tables.table_name = columns.table_name
WHERE tables.table_type = 'base table'
AND columns.data_type = 'float'
ORDER BY columns.table_name, columns.column_name

Can we look into LOG file?

2007-05-06T04:01:00.000-07:00

In SQL Server 2005 with fn_dblog function you can get more info about what's happening in SQL Server. MS has added a new data like 'partitionid' column for example.

Let us do some testing

USE Demo
--create a table for testing
CREATE TABLE dbo.Test1 (c INT)
--populate the table with data

WITH numbers AS
(SELECT 1 as Num
UNION ALL
SELECT Num + 1 FROM numbers WHERE Num <10)
INSERT INTO dbo.Test1 SELECT Num FROM numbers OPTION(Maxrecursion 10)

--delete the data
DELETE FROM dbo.Test1

In order to identify how many 'DELETE' operation and when does it start? Run the followinq query.

SELECT [Transaction Name],[begin time FROM ::fn_dblog( NULL,NULL)WHERE [Transaction Name] ='DELETE'

Ok, the next step I'd like to know how long does DELETION take? So we have [end time column] for this purpose. We need to write the below query which gets little complicated.

WITH cte_delete
AS
(
SELECT [Transaction Name],[begin time],[end time]
FROM ::fn_dblog( NULL,NULL)d
WHERE [begin time] IS NOT NULL OR [end time]IS NOT NULL
) ,cte_d
AS
(
SELECT [Transaction Name],[begin time],(SELECT MIN([end time]) FROM cte_delete C
WHERE cte_delete.[begin time]<=C.[end time])AS [end time]
FROM cte_delete WHERE [Transaction Name] ='DELETE' OR
[Transaction Name] IS NULL
)
SELECT * FROM cte_d
WHERE [begin time] IS NOT NULL OR [end time]IS NOT NULL

Also take a look at 'Operation' column.

SELECT Operation,AllocUnitName,[Transaction Name],[begin time],[end time]
FROM ::fn_dblog( NULL,NULL)
WHERE Operation ='LOP_DELETE_ROWS' OR [Transaction Name]='DELETE'

To identify which tables are being hit by DELETE operation.

SELECT AllocUnitName, COUNT([Current LSN])
FROM ::fn_dblog(NULL,NULL )
WHERE Operation = N'LOP_DELETE_ROWS'
GROUP BY AllocUnitName

String manipulation (extracting)

2007-05-01T01:27:00.000-07:00

I probably should create some group of articles for string manuipulation.At this time I'd like to show you an example to extract portion of text prior/after to some charcter/symbol.An idea is simple to iterate through string and return the text prior (in my example) to 'N' appearance of specified symbol.

CREATE FUNCTION dbo.udf_Extract(
@target varchar(8000),
@string varchar(8000),
@i int
) RETURNS varchar(8000) AS BEGIN
DECLARE @pos int
SET @pos = 0
WHILE @i > 0 BEGIN
SET @pos = CHARINDEX(@target,@string,@pos+1)
SET @i = @i - 1
IF @pos = 0 RETURN '0'
END
RETURN SUBSTRING (@string,1,@pos-1)
END
GO

--Usage
SELECT dbo.udf_Extract('@','hjdhjdhj@hjjsd@hjdfhdfj@jjj,3)

Which tables have a trigger/s?

2007-04-29T03:44:00.000-07:00

I have been recently helping my fried on very large database ,which has hundred od tables to determine which table has a trigger. So I came out with the below script that may help others to identify triggers define on the tables.

--SQL Server 2000(SP3)

SELECT
name
, CASE WHEN instrig <> 0 then OBJECT_NAME (instrig) ELSE 'None' END as
'Insert Trigger'
, CASE WHEN updtrig <> 0 then OBJECT_NAME (updtrig) ELSE 'None' END as
'Update Trigger'
, CASE WHEN deltrig <> 0 then OBJECT_NAME (deltrig) ELSE 'None' END as
'Delete Trigger'
FROM
sysobjects
WHERE
type = 'U'
AND
( instrig <> 0
OR
updtrig <> 0
OR
deltrig <> 0
)

What's the version of SQL Server?

2007-04-22T02:48:00.000-07:00

I have seen this question (see post's subject) so many times in newsgroup asking people, so I decided to post some usefull info about SQL Server by using SERVERPROPERTY command.

SELECT @@VERSION AS [Server Information]

SELECT SERVERPROPERTY('ServerName') AS [ServerName]
SELECT SERVERPROPERTY('MachineName') AS [MachineName]
SELECT SERVERPROPERTY('InstanceName') AS [InstanceName]

SELECT SERVERPROPERTY('productversion') AS [ProductVersion]
SELECT SERVERPROPERTY('ProductLevel') AS [ProductLevel]
SELECT SERVERPROPERTY('Edition') AS [Edition]

SELECT SERVERPROPERTY('Collation') AS [Collation]
SELECT SERVERPROPERTY('IsClustered') AS [IsClustered]
SELECT SERVERPROPERTY('IsFullTextInstalled') AS [IsFullTextInstalled]
SELECT SERVERPROPERTY('IsIntegratedSecurityOnly') AS
[IsIntegratedSecurityOnly]
SELECT SERVERPROPERTY('LicenseType') AS [LicenseType]
SELECT SERVERPROPERTY('NumLicenses') AS [NumLicenses]

Mixed or Windows Authentication Only

2007-04-19T03:44:00.000-07:00

Ok, at this time I would to discuss about the one but realy painful issue we have to face where we use SQL Authentication (Mixed).
Consider the following situation. You are using SQL Server (I speak for SQL Server 2000 SP4) Authentication(Mixed).
You also have a database and user within which is mapped to the SQL login. Now , you have been told to move the database on the new server. So you created a backup of the database , copied it to the new server as well as create a SQL Login with the same name as you had on the old one. Ok, now it is time to perform restore database, so everything went just fine but when theh users tried to connnect to the database (through application) on the new server they get the error. What happened? The thing is that SQL Server is 'losing' SIDs that mapped between a login and user when you created a new SQL Login on the new server (master..syslogins) and a database user (dbname..sysuser)that you brought within a database. So we have sp_change_users_login system stored procedure that takes care of re-mapping SID's.

However, that DOES NOT happen if you use Windows Authentication Only.That means SQL Server keeps SID's realtionship between Windows Login and User database although you move the database on the new server.

Just one more argument to use Windows Authentication Only.

In SQL Server 2005 SP2 MS has introduced some greate features on subject.
http://blogs.msdn.com/lcris/archive/2007/02/19/sql-server-2005-some-new-security-features-in-sp2.aspx.

Drop all indexes

2007-04-15T02:13:00.000-07:00

Well, sometimes we have to do that. A large INSERT statement may run faster for instance. The below script just generate DROP command of all indexes in current dataabse. You can add a filter to specify a table name for example.(SQL Server 2000)

SELECT
'DROP INDEX ' +
QUOTENAME(USER_NAME(o.uid)) +
'.' +
QUOTENAME(o.name) +
'.' +
QUOTENAME(i.name)
FROM sysobjects o
JOIN sysindexes i ON
i.id = o.id
WHERE i.indid BETWEEN 1 AND 254 AND
INDEXPROPERTY(i.id, i.name, 'IsStatistics') = 0 AND
INDEXPROPERTY(i.id, i.name, 'IsHypothetical') = 0 AND
OBJECTPROPERTY(o.id, 'IsMSShipped') = 0

Some third parties break DIFFERENTIAL backup chains

2007-04-10T00:54:00.000-07:00

I'd like to share some things I have seen on my client's machine. He has a full backup database set up once a night and every four hours differential back during the work day. But at the same time ( it should be done at night) the network administrator has ran tird party software backup to copy .MDF/LDF files. When we started to restore database it throws the error that SQL Server cannot apply this bakup file to the full backup file created at night. We did lots of testing on the developing machine and did not het the error. And only when we asked the net.admin not to run this software then we were able to restore the database succefully.

Note, I have not seen that doing LOG files backup does break the chains.

So please do lots of testing before you find your self without properly created disaster recovery strategy.

I found that MS has this KB explaining the behaviour.
http://support.microsoft.com/kb/903643.

Output clause and Auditing of data

2007-03-26T05:23:00.001-07:00

There is very common scenario to auditing of data in a table using triggers. In this case, the trigger uses information from the inserted and updated tables to add rows into the audit tables. MS has introduced a new OUTPUT clause in SQL Server 2005 to audit data instead of using triggers to insert rows into an audit table.

CREATE TABLE test (col INT NOT NULL);
CREATE TABLE test_audit ( old_col INT NOT NULL, new_col INT NULL);
INSERT INTO test (col) values( 1 );
INSERT INTO test (col) values( 2 );

UPDATE test
SET col = col + 1
OUTPUT deleted.col ,inserted.col into test_audit
WHERE col = 1;

DELETE FROM test
OUTPUT deleted.col, NULL into test_audit
WHERE col = 2;

SELECT * FROM test
SELECT * FROM test_audit;

Detaching and Attaching database

2007-03-21T01:04:00.000-07:00

At this time I'd like to share some scripts that used undocumented (sorry) but very popular among DBA's sp_MSforeachdb stored procedure. If you have hundred of databases on your server and you was told to move those databases to another server the quick way to do the job is performing the following scripts. First of all we detach all user databases from the server and the second query will output an attach
script. The thing is that the second query will output the script that attach the database that has more than one LOG file.

I'm sure that you have backuped the user databases before performing such tasks.

--Detaching
EXEC sp_MSforeachdb
'IF N''?'' NOT IN(''model'' ,''master'', ''msdb'', ''tempdb'')
BEGIN
SELECT ''EXEC sp_detach_db ''''?'''',''''true''''''
END'

--Attaching
EXEC sp_msforeachdb
'SELECT ''EXEC sp_attach_db ''''?'''',
'''''' + RTRIM(filename) + '''''',''
FROM ?..sysfiles
WHERE fileid = (SELECT MIN(fileid) FROM ?..sysfiles)
UNION ALL
SELECT '''''''' + RTRIM(filename) + ''''''''
FROM ?..sysfiles
WHERE fileid > (SELECT MIN(fileid) FROM ?..sysfiles) AND
fileid < (SELECT MAX(fileid) FROM ?..sysfiles)
UNION ALL
SELECT '''''''' + RTRIM(filename) + ''''''''
FROM ?..sysfiles
WHERE fileid = (SELECT MAX(fileid) FROM ?..sysfiles)'

String manipulation (searching)

2007-03-14T00:39:00.000-07:00

Recently I have seen some post in the news group regarding seraching a string in the table or even part of string. Yes , there are lots of resources in the internet how to do that. I would like to shere some ideas also here.

Consider the following structure. We need to find in col1 a part of string. Let us find occurence of '23'.

CREATE TABLE #Test
(
col INT,
col1 VARCHAR(100)
)
INSERT INTO #Test VALUES (1,'12,23,44,56')
INSERT INTO #Test VALUES (2,'26,45,22,32')
INSERT INTO #Test VALUES (3,'45,22,34,23')

DECLARE @st VARCHAR(10)
SET @st='23'
SELECT *
FROM #test
WHERE ','+col1+',' LIKE '%,'+@st+',%'

Now, take a look at following table
CREATE TABLE #Test (contact VARCHAR(20) PRIMARY KEY)

INSERT INTO #Test VALUES ('Bill Clinton')
INSERT INTO #Test VALUES ('George W. Bush')
INSERT INTO #Test VALUES ('John F Kennedy')
INSERT INTO #Test VALUES ('Smith')

1)
SELECT contact
FROM #Test
WHERE contact LIKE '%[A-Z] [A-Z][^A-Z]%[A-Z]'
----
2)
SELECT contact
FROM #Test
WHERE contact LIKE '% _ %' or contact LIKE '% _. %'

The first one returns all people that have a middle name. The second one returns
people who has not '.' in this name after middle name.

Script Objects SS2K

2007-03-11T05:05:00.000-07:00

Well, there are still lots of people/businesses use SQL Server 2000 I'd like to put the following script using SQL DMO object library. You have to open a new project with VB6 and put the code in the new form.
The cool thing is that you get the script without default COLLATION to be specify.
See if you need to run the script on another server/s or database/s that have different collation.

Also, Tibor Karaszi has great info at his web site
http://www.karaszi.com/SQLServer/info_generate_script.asp

Sub ScriptDB(strLogin As String, strPwd As String, _
strDataBase As String, StrFilePath As String)

Dim sql As Object
Dim db As Object
Dim objTrigger As Object
Dim intOptions As Long
Dim genObj

Set sql = CreateObject("SQLDMO.SQLServer")
Set db = CreateObject("SQLDMO.Database")
Set objTrigger = CreateObject("SQLDMO.Trigger")

Const sDrops As Integer = 1
Const sIncludeHeaders As Long = 131072
Const sDefault As Integer = 4
Const sAppendToFile As Integer = 256
Const sBindings As Integer = 128
' No collation to be specify
Const SQLDMOScript2_NoCollation As Long = 8388608

' We need to specify multiple behaviors
' for the ScriptType argument, you use "Or" to combine these.
intOptions = sDrops Or sIncludeHeaders Or _
sDefault Or sAppendToFile Or sBindings Or SQLDMOScript2_NoCollation

' Connect to server
sql.Connect "(local)", strLogin, strPwd
Set db = sql.Databases(strDataBase, "dbo")

' Script User Defined Data Types
For Each genObj In db.UserDefinedDatatypes
genObj.Script intOptions, StrFilePath
Next

' Script Tables and Triggers, ignoring system
' tables and system generated triggers
For Each genObj In db.Tables
If genObj.SystemObject = False Then
genObj.Script intOptions, StrFilePath,,SQLDMOScript2_NoCollation
For Each objTrigger In genObj.Triggers
If objTrigger.SystemObject = False Then
objTrigger.Script intOptions, StrFilePath
End If
Next
End If
Next

' Script Rules
For Each genObj In db.Rules
genObj.Script intOptions, StrFilePath
Next

' Script Defaults
For Each genObj In db.Defaults
genObj.Script intOptions, StrFilePath
Next

' Script Sprocs, ignoring system sprocs
For Each genObj In db.StoredProcedures
If genObj.SystemObject = False Then
genObj.Script intOptions, StrFilePath
End If
Next

' Script Views, ignoring system views and informational schemas
For Each genObj In db.Views
If genObj.SystemObject = False Then
genObj.Script intOptions, StrFilePath
End If
Next

End Sub

To call the procedure save the module and supply the parameters.
Call ScriptDB("UserName","Password","DatabaseName","C:\DBScript.SQL")

Getting the value of Constraint

2007-03-11T04:58:00.000-07:00

Recently I was asked by my colleague to return tha actual value of CHECK constaint that he has defined on the table.
I think it can be usefull for some reports that we've done on the database.
Tested on SQL Server 2000(SP3a) and SQL Server 2005(SP2a)

CREATE TABLE Test
(
col INT,
coldatetime DATETIME
)

ALTER TABLE Test
ADD CONSTRAINT coldatetime_check
CHECK ( coldatetime > '20070101')

SELECT text FROM syscomments
WHERE id =OBJECT_ID('coldatetime_check')