Dear Pinal,

I am new to SQL Server and our regular DBA is on vacation. Our production database had some problem and I have just restored full database backup to production server. When I try to apply log back I am getting following error. I am sure, this is valid log backup file. Screenshot is attached.

[Few other details regarding server/ip address removed]

Msg 3117, Level 16, State 1, Line 1
The log or differential backup cannot be restored because no files are ready to roll forward.
Msg 3013, Level 16, State 1, Line 1
RESTORE LOG is terminating abnormally.

Screenshot attached. [Removed as it contained live IP address]

Please help immediately.

Well I have answered this question in my earlier post, 2 years ago, over here SQL SERVER – Fix : Error : Msg 3117, Level 16, State 4 The log or differential backup cannot be restored because no files are ready to rollforward. However, I will try to explain it a little more this time.

For SQL Server database to be used it should in online state. There are multiple states of SQL Server Database.

• ONLINE (Available – online for data)
• OFFLINE
• RESTORING
• RECOVERING
• RECOVERY PENDING
• SUSPECT
• EMERGENCY (Limited Availability)

If the database is online, it means it is active and in operational mode. It will not make sense to apply further log from backup if the operations have continued on this database. The common practice during the backup restore process is to specify the keyword RECOVERY when the database is restored. When RECOVERY keyword is specified, the SQL Server brings back the database online and will not accept any further log backups.

However, if you want to restore more than one backup files, i.e. after restoring the full back up if you want to apply further differential or log backup you cannot do that when database is online and already active. You need to have your database in the state where it can further accept the backup data and not the online data request. If the SQL Server is online and also accepts database backup file, then there can be data inconsistency. This is the reason that when there are more than one database backup files to be restored, one has to restore the database with NO RECOVERY keyword in the RESTORE operation.

I suggest you all to read one more post written by me earlier. In this post, I explained the time line with image and graphic SQL SERVER – Backup Timeline and Understanding of Database Restore Process in Full Recovery Model.

Sample Code for reference:

RESTORE DATABASE AdventureWorks
FROM DISK = 'C:\AdventureWorksFull.bak'
WITH NORECOVERY;
RESTORE DATABASE AdventureWorks
FROM DISK = 'C:\AdventureWorksDiff.bak'
WITH RECOVERY;

In this post, I am not trying to cover complete backup and recovery. I am just attempting to address one type of error and its resolution. Please test these scenarios on the development server. Playing with live database backup and recovery is always very crucial and needs to be properly planned. Leave a comment here if you need help with this subject.

Similar Post:
SQL SERVER – Restore Sequence and Understanding NORECOVERY and RECOVERY

Note: We will cover Standby Server maintenance and Recovery in another blog post and it is intentionally, not covered this post.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

Here, I need your help. Please your comments, suggestions, expectation or potential interview Question (along with answer) here. Your input will be very valuable. As time goes by we all learn and get better. There were few things missing at that time when those interview questions and answers were prepared, now is the time to complete the gap and make this interview questions more useful.

If you know all, this Question and Answers are not for you. This are for those who is eager to learn and need help in the area.

If you do not want to leave a comment, I suggest to send me email at pinal “at” SQLAuthority.com

Following is the reproduction of original consolidation post for quick reference.

SQL SERVER – 2008 – Interview Questions and Answers – Part 1

SQL SERVER – 2008 – Interview Questions and Answers – Part 2

SQL SERVER – 2008 – Interview Questions and Answers – Part 3

SQL SERVER – 2008 – Interview Questions and Answers – Part 4

SQL SERVER – 2008 – Interview Questions and Answers – Part 5

SQL SERVER – 2008 – Interview Questions and Answers – Part 6

SQL SERVER – 2008 – Interview Questions and Answers – Part 7

SQL SERVER – 2008 – Interview Questions and Answers – Part 8

Download SQL Server 2008 Interview Questions and Answers Complete List

Reference : Pinal Dave (http://blog.SQLAuthority.com)

Let us consider the following example of this query.

The following query usually runs on multicore on a dual core machine (please note it may not be the case with your machine).

USE AdventureWorks
GO
SELECT *
FROM Sales.SalesOrderDetail
ORDER BY ProductID
GO

Now the same query can be ran on a single core with the usage of MAXDOP query hint. Let us see the query for the same.

USE AdventureWorks
GO
SELECT *
FROM Sales.SalesOrderDetail
ORDER BY ProductID
OPTION (MAXDOP 1)
GO

Execution plan for the query with query hint of maxdop (1) does not have parallelism operator in it. This way we can remove the parallelism using MAXDOP.

However, before playing with this query hint, please make sure that you check your performance using an execution plan. It is quite possible that the performance of Query with MAXDOP as query hint may be quite degraded when compared to the original performance. You should be very careful with this hint.

Let us compare in our case what is the performance difference between the two above queries. The difference between those two queries is only the query hint of MAXDOP.

In our example, we got degraded performance as we restricted our query on a single CPU. This is not necessary in the case of all the queries. MAXDOP may improve or reduce performance, test your query out.

I have now one question for all readers. Do you use this query hint? If you do use it, then what is the purpose of the same. Please leave a comment here.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

Question 1: I have a unique requirement where I do not want to use any index of the table; how can I achieve this?

Question 2: Currently my table uses clustered index and does seek operation; how can I convert seek to scan?

First of all, I am not going to analysis their need of why, in fact, they want to convert seek to scan or use no index here. The requirement is strange as using no index or scanning large table may reduce the performance instead of improving it. We are not going to discuss the ‘It Depends’ condition where scan can be better than seek or seek can be better than scan. The User has to take his/her own decision after looking at performance which one is better for its need.

Now, let us see how we can force the use of no index or force a scan operation on clustered index. Scan operation on clustered index is, in fact, scanning the original table. When querying table sys.indexes and closely looking at the column type, it is found to contain various values. Value 0 (Zero) in column type indicates that it is a heap. A heap is defined as a table without index. We will use the value 0 on the table with index to simulate the effect of no index or force index scan where index seek is happening.

If any object has a clustered or nonclustered index on it, it will not have entry of index_id =0. The value 0(Zero) for index_id indicates the heap. You can find further details at sys.indexes.

Now, let us run the following query, and please note the usage of the Index hint of INDEX(0), it will force query to not to use any index or behave like a query on heap. The value which is passed to the function INDEX is the value of index_id.

USE AdventureWorks
GO
SELECT ProductID, ProductNumber, Color
FROM Production.Product
WHERE ProductID < 500
GO
SELECT ProductID, ProductNumber, Color
FROM Production.Product WITH (INDEX(0))
WHERE ProductID < 500
GO

Let us check the execution plan for this query batch, and we should able to see in the second query that the seek is converted to scan. This answers the question of how to convert seek to scan?

Now, once again, let us run the following similar query, where we can see how INDEX(0) is behaving.

USE AdventureWorks
GO
SELECT ProductID, ProductNumber
FROM Production.Product
WHERE ProductNumber = 'CR-7833'
GO
SELECT ProductID, ProductNumber
FROM Production.Product WITH (INDEX(0))
WHERE ProductNumber = 'CR-7833'
GO

It is clear from result that original index usage is replaced with Clustered Index Scan. Also, from the first example, we have seen that how it is forcing scan even though seek is possible. Well, this indicates that the basic index was not used. I would rather say that we get a result that is same as the result when there is no index on the table.

The strong argument here is that even though you use Index(0) or Index(1) in this later example, there is no difference in the result. The logical reasoning we can reach after considering the first example is that if the use of Index(1) was forcing index seek due to any reason, using Index(0) will eliminate the seek part and clustered index will be scanned just like it would have happened in the case of Heap. If table has clustered index, it is ordered in a certain way and it cannot be used like heap after having clustered index on it; however just like heap, the scan is pushed over it.

I want to know if community had ever faced the above two questions and also your take on these. Moreover, I have a gut feeling that there is something missing in the above explanation, and I request you all to help me complete the explanation.

Reference: Pinal Dave (http://blog.SQLAuthority.com)

Official site: What’s New in SQL Server 2008 R2 Editions

However, I have bookmarked following site of Brent Ozar SQL Server 2008 R2 Pricing and Feature Changes. I think Brent has answered one very interesting question there that SQL Server R2 is FREE for only who has purchased software assurance.

I really would like to know if your organization is upgrading to SQL Server 2008 R2. Please leave your comment here.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

The most interesting comment conversation is among Divya, Brian and Marko. Please read the comments of Marko for sure. It is the comment, which has triggered this post.

Comments by Divya

I have seen in one of the blogs to use EXISTS like

IF EXISTS(Select null from table)

Will it optimize the perfomance better than

IF EXISTS(Select 1 from table)?

Comments by Brian Tkatch

@Divya

It makes no difference what is put there.

Brad Schulz has an interest article on it: http://bradsruminations.blogspot.com/2009/09/age-old-select-vs-select-1-debate.html

Even 1/0 is allowed! Obviously, it is not evaluated.

Technically, however, * does get expanded, adding some minuscule amount of time: http://www.sqlskills.com/BLOGS/CONOR/2008/02/default.aspx?page=2

So, anything other than * takes the same amount of time. * takes an iota longer. Personally, i use *, to show that i do not care what the results are.

Comments by Pinal Dave

Hello Divya,

I not think so because there is no significant difference in transferring of 1 byte and a null value.

Other than that there is no difference between these two queries.

Regards,

Pinal Dave

Comments by Marko Parkkola

“I not think so because there is no significant difference in transferring of 1 byte and a null value.”

While that is true in this exact case I like to be pedantic about this and elaborate this a bit and probably confuse everybody up :)

There is no concept NULL value if you think about C code or even the CPU. Pointers in C can have NULL values but that just means that the pointer is pointing to memory in the address 0×00000000 (in 32-bit machine). The actual pointer is still taking sizeof(VOID*) amount of memory even when it points to NULL.

Now if you think about functions in C or any other language they always reserve space in the function stack for the return value. You can say, again in C, that you don’t care about the return value and declare the function to return “void” but still it takes 4 bytes of memory in 32-bit machine for return value.

And actually, of you think about the CPU, there is 32 bit register reserved exactly for this.

And now considering that it makes no difference to return one byte or four bytes since there is space for four bytes anyway. And in fact, many times (all the times?) when you deal with C data types CHAR (1 byte) or SHORT (2 bytes) you end up taking 4 bytes because of the padding to keep memory aligned.

Uh! I almost felt like going back in time some ten years when I was writing kernel drivers for Windows :)

This is excellent participation from experts like Marko and Brian. I really like the last line of Marko where it mentioned kernel drivers of Windows. Bravo everybody!

If you have any suggestion please leave a comment in original article.

Reference: Pinal Dave (http://blog.SQLAuthority.com)

I have quite often seen that information_schema.routines gives proper dependency relationship where assp_depends returns an incorrect answer. However, as per book online sp_depends will be deprecated, and instead, sys.dm_sql_referencing_entities and sys.dm_sql_referenced_entities are recommended.

Let us quickly see where sp_depends fail and other solutions work fine.

Let us first create two scenarios.

Scenario 1: Normal Table Creation Order, where objects are created first and then used afterwords.

USE TempDB
GO
CREATE TABLE dbo.TestTable
( ID INT,
Name VARCHAR(100))
GO
-- dbo.First is not created yet
CREATE PROCEDURE dbo.Second
AS
EXEC dbo.First
GO
CREATE PROCEDURE dbo.First
AS
SELECT ID, Name
FROM TestTable
GO


Scenario 2: Objects are created afterwords and they are referenced first.

USE TempDB
GO
CREATE TABLE dbo.TestTable
( ID INT,
Name VARCHAR(100))
GO
CREATE PROCEDURE dbo.First
AS
SELECT ID, Name
FROM TestTable
GO
-- dbo.First is already created
CREATE PROCEDURE dbo.Second
AS
EXEC dbo.First
GO

Now let us run following three queries on both the scenarios.

-- Method 1: Using sp_depends
sp_depends 'dbo.First'
GO
-- Method 2: Using information_schema.routines
SELECT *
FROM information_schema.routines ISR
WHERE CHARINDEX('dbo.First', ISR.ROUTINE_DEFINITION) > 0
GO
-- Method 3: Using DMV sys.dm_sql_referencing_entities
SELECT referencing_schema_name, referencing_entity_name,
referencing_id, referencing_class_desc, is_caller_dependent
FROM sys.dm_sql_referencing_entities ('dbo.First', 'OBJECT');
GO

Result from Scenario 1

Result from Scenario 2

It is clear that sp_depends does not give proper/correct results when the object creation order is different or following deferred name resolution.

I suggest the use of the third method, in which sys.dm_sql_referencing_entities is used.

Use the following script to get correct dependency:

SELECT referencing_schema_name, referencing_entity_name,
referencing_id, referencing_class_desc, is_caller_dependent
FROM sys.dm_sql_referencing_entities ('YourObject', 'OBJECT');
GO

Let me know the type of scripts you use for finding Object Dependencies. I will post your script with due credit.

Reference: Pinal Dave (http://blog.SQLAuthority.com)

Reference: Pinal Dave (http://blog.SQLAuthority.com)

Q. How the records of a table can be scripted in INSERT INTO statements?

A. In SQL Server 2008 : Right click Database > Tasks > Generate Scripts > In the wizard on Choose Script Option page, set Script Data option to True and complete the wizard.For SQL 2005 or earlier versions, use Database Publishing Wizard. For more details about Database Publishing wizard, please visit the blog http://blog.sqlauthority.com/2007/11/16/sql-server-2005-generate-script-with-data-from-database-database-publishing-wizard/

Q. How can I track the changes or identify the latest insert-update-delete from a table?

A. In SQL Server 2005 and earlier versions, there is no inbuilt functionality to know which row was recently changed and what were the changes. However, in SQL Server 2008, a new feature known as Change Data Capture (CDC) has been introduced to capture the changed data. To get more details about CDC, please visit http://blog.sqlauthority.com/2009/08/15/sql-server-introduction-to-change-data-capture-cdc-in-sql-server-2008/

In versions prior to 2008, the best method to track changes at the table level is Triggers and at the database level is SQL Profiler. Another workaround is to add one more column in the table like “ModifiedDate” to keep the current date-time of the change. Timestamp, rowversion and checksum columns are for optimistic concurrency. These are not suitable to track changes later.

Q. I am getting collation conflict error when joining columns of two tables?

A. When we try to compare, join or union two columns of different collations, the following error occurs: “Cannot resolve the collation conflict for equal to operation”.

Collation represents the character set of a data and can be specified at server, database and column levels. Collation of server is default for databases and collation of database is default for all columns. The collation is matched in following cases while executing a SELECT statement:

SELECT col1, col2
FROM Table1.col1 = Table2.col4
UNION
SELECT col5, col6
FROM Table3

Here make sure that collation of following columns should match
col1 = col5
col1 = col4
col2 = col6

To resolve this issue, you can either specify collation with column name in t-sql statement or change the collation of column permanently by altering the table, database or even server. For more details on how to specify collation in t-sql statement, please check following article: http://blog.sqlauthority.com/2007/06/11/sql-server-cannot-resolve-collation-conflict-for-equal-to-operation/

Sometimes, this error occurs while comparing the table with temporary tables. In this case, check the default collation of your working database and tempdb. If they do not match, then apply the above mentioned resolution.

Q. How can I get data from a database on another server?

A. If you want to import data only through t-sql query, then either use OPENDATASOURCE function. To repeatedly get data from another server, create a linked server and then use OPENQUERY function or use 4 part naming. If you are not adhered with T-SQL, then better use import/export wizard, and you can save it as a SSIS package for future use.

Q. My multilanguage data is not inserting correctly and returning as ???. How can I store Multilanguage data in a table?

A. There are two things to keep in mind while storing unicode data. First, the column must be of unicode data type (nchar, nvarchar, ntext). Second, the value must be prefixed with N while insertion. For example,

INSERT INTO table (Hindi_col) values (N’hindi data’)

I will post the second series of readers’ question and answers in future posts.

Reference: Pinal Dave (http://blog.SQLAuthority.com)

SQL SERVER – Size of Index Table – A Puzzle to Find Index Size for Each Index on Table

The puzzle was to write a query that will return the size for each index that is on any particular table. We need a query that will return an additional column in the above listed query and it should contain the size of the index.

So far I have found two potential solutions. I have done some good amount of testing of both the script and they both are really giving accurate results 99.9% of the time. There were couple of instances where it provided incorrect results but for the most of the time, it just worked like charm. Again, it may be a bit of rounding up logic and I will not hold it against the script.

I want to thank Gaurav Sharama and Adam Hutson for their participation.

Solution from Gaurav Sharma

DECLARE @OBJECT_NAME VARCHAR(255) = 'HumanResources.Shift';
DECLARE @temp TABLE
(
indexID  BIGINT,
objectId  BIGINT,
index_name NVARCHAR(MAX),
used_page_count  BIGINT,
pages  BIGINT
)
--Insert into temp table
INSERT INTO @temp
SELECT
P.index_id,
P.OBJECT_ID ,
I.name,
SUM (used_page_count),
SUM (
CASE
WHEN (p.index_id < 2) THEN (in_row_data_page_count + lob_used_page_count + row_overflow_used_page_count)
ELSE lob_used_page_count + row_overflow_used_page_count
END
)
FROM sys.dm_db_partition_stats P INNER JOIN sys.indexes I ON I.index_id = P.index_id AND I.OBJECT_ID = P.OBJECT_ID
WHERE p.OBJECT_ID = OBJECT_ID(@OBJECT_NAME)
GROUP BY P.index_id, I.Name, P.OBJECT_ID;
SELECT index_name INDEX_NAME,
LTRIM (STR ((CASE WHEN used_page_count > pages THEN (used_page_count - pages) ELSE 0 END) * 8, 15, 0) + ' KB') INDEX_SIZE
FROM @temp T
GO

Solution from Adam Hutson

DECLARE
@objname NVARCHAR(776),
@id INT,
@dbname sysname
SELECT
@objname = 'HumanResources.Shift',
@dbname = ISNULL(PARSENAME(@objname, 3),DB_NAME()),
@id = OBJECT_ID(@objname)
SELECT
i.*,
CASE
WHEN ps.usedpages > ps.pages THEN (ps.usedpages - ps.pages)
ELSE 0
END * 8 indexsize
FROM sys.indexes i
INNER JOIN (
SELECT
OBJECT_ID,
index_id,
SUM (used_page_count) usedpages,
SUM (
CASE
WHEN (index_id < 2) THEN (in_row_data_page_count + lob_used_page_count + row_overflow_used_page_count)
ELSE lob_used_page_count + row_overflow_used_page_count
END
)pages
FROM sys.dm_db_partition_stats
WHERE OBJECT_ID = @id
GROUP BY OBJECT_ID, index_id
) ps ON i.index_id = ps.index_id
WHERE i.OBJECT_ID = @id
GO

If you have different script or have suggestions on this script, please let us know.

Reference: Pinal Dave (http://blog.sqlauthority.com)

USE AdventureWorks
GO
EXEC sp_spaceused [HumanResources.Shift]
GO

Above query will return following resultset

The above SP provides basic details such as rows, data size in table, and Index size of all the indexes on the table.

If we look at this carefully, a total of three indexes can be found on the table HumanResources.Shift.

USE AdventureWorks
GO
SELECT *
FROM sys.indexes
WHERE OBJECT_ID = OBJECT_ID('HumanResources.Shift')
GO

The above query will give result with query listing all the index on the table.

There is a small puzzle for all of you here. The puzzle is to write a query that will return the size for each index that is listed in above query. We need a query that will return an additional column in the above listed query and it should contain the size of the index. In our case, we will have three different sizes, which should add up to a total of 40 KB as shown in earlier query, where the total size is displayed.

I will publish the solution with due credit on this blog.

Reference: Pinal Dave (http://blog.sqlauthority.com)

Hats off to you Brian! You can reach Brian at his email address (brians.sql.blog (at) gmail (dot) com) or leave comment here.

Download all scripts and explanation here

About Collation Consolidation

At some time in your DBA career, you may find yourself in a position when you sit back and realize that your database collations have somehow run amuck, or are faced with the ever annoying CANNOT RESOLVE COLLATION message when trying to join data of varying collation settings.  I’m sure there are as many reasons as to how you’ve come to this juncture as there are readers.  Maybe you have DBA’s and instances around the world and the local staff are using their own collations.  Or, maybe you’ve learned the hard way that ALTER DATABASE x SET COLLATION y doesn’t traverse.  Whatever it is, you’re probably here for the one underlying reason.  Manually consolidating collations would be a nightmare.  So, without loading to much fluff, let’s jump right in.

Here is the basic approach:

1. Identify all the columns which have incorrect collation values.
2. Identify all objects bound to above columns, including indexes, keys, checks and UDF’s,
3. Script out all of the above object DDL.
4. Drop all of the objects.
5. Update all the columns with the correct collation
6. Recreate all of the dropped objects using the DDL generated in step 3

This is done with a collection of non-destructive scripts which generate the DDL to perform all of the above.  This generated DDL MUST be saved to scripts for later use on the target databases.  And obviously, you’ll need one set of scripts per database.  If that requires clarification, just stop reading now.

In this article, I give to you the four scripts to do all this and will try to provide as much understanding as possible.  The first script is purely to create a stored procedure which will be repeatedly called.  This proc will need to exist in a database named “DBADB”.  If you want to use a different database, you’ll need to do a search and replace in the first script and the last script since the database name is hard-coded.  I tried to make these somewhat generic, but the only approach that came to mind would be to use OLE to invoke the VBScript engine to prompt you for the name of the database where this stored procedure is located.  Overkill in my opinion.  Ctrl-H is your friend here.

The script to perform step 3 above uses DMO via OLE, so if you’ve disabled the extended stored procedures; sp_OACreate, sp_OASetProperty, sp_OAMethod, etc, you won’t be able to continue until they’ve been enabled.

One final note.  If you’re not a member of the System Administrators role, please don’t try to execute these.  You’ll probably find your account revoked quickly.

Ok, time to roll up your sleeves.

Script 1 – The DDL Generator

This first script is to create a stored procedure.  This procedure must reside in a location that will not interfere with other databases, whether they’re development, test or production.  The SP uses DMO via OLE to capture a buffer dump of the object code.  The result is the same as if you right-clicked an object and selected “Script to new window as…”

The DMO hierarchy of a SQL instance is weird.  Object reference isn’t particularly intuitive.  But, for the most part, objects are contained within the associated table collection.  In DBA language, that would look like

Table 1:M Checks

Table 1:M Keys

Table 1:M Indexes

Database 1:M UserDefinedFunctions

Therefore, to locate a particular object in the DMO hierarchy, you would need to pass in all the required parameters, including the object type.  So, for checks, keys and indexes, you must also pass in the database name and the parent table name.  For functions, only the database name.

sp_scriptObject(

@in_dbName SysName,

@in_objType VarChar(100),

@in_objName VarChar(100),

@in_objSubName VarChar(100) = NULL

)

Now, the good news.  The subsequent scripts will do all of the above automatically, and this script will be called repeatedly.

Implementing this script.

As I mentioned, this script will be called repeatedly.  The script from whence it’s called will need to be modified to make reference to the correct database name.  These scripts have been designed to work with a DBA-use only database called “DBADB”.  Therefore, if you choose to keep this procedure located in a database of a different name, you’ll need to perform a search and replace for all the “DBADB” reference.  Otherwise, you will need to create a database named “DBADB”.

Execute the first script.  This will create the necessary stored procedure.

Script 2 – Drop Dependencies

The name for this script is somewhat deceiving.  It is a non-destructive script and only performs a set of SELECT statements.  Rather, this script generates a collection of DROP [object] statements.

The goal here is to identify all the objects that are bound to columns which have incorrect collation settings.  There is no clean way to pull them all back in a single statement since determining this is handled differently for the various types of objects.  So, they’re all concatenated together with UNION’s.  The orders in which these are executed are as follows:

1. Table Functions
2. Unique Constraints
3. Check Constraints
4. Indexes

You must also edit this script to use your desired collation.  Perform a search and replace of SQL_Latin1_General_CI_AS to your chosen collation.

It’s performed in this order in attempt to avoid dependency issues.

The result is a collection of DROP statements.  This must then be copy and pasted into a script and saved for later use.  Best to use a naming convention that reflects the name of the database, and that this will be the first script to execute when you’re ready to begin the patching, for example: myDatabase_post-script_1.sql, myDatabase_post-script_2.sql, etc.

You should take note, it is possible that some databases may not have any objects that are bound to columns that need to be updated, so it’s perfectly fine if this script produces no results.

Script 3 – Change Collations

Much like script 2, this script is also non-destructive and will produce a collection of DDL statements to update the column collations.

You must also edit this script to use your desired collation.  Perform a search and replace of SQL_Latin1_General_CI_AS to your chosen collation.

VERY IMPORTANT.  These scripts were originally intended to allow us to consolidate all our collations in a SQL 2000 to SQL 2008 migration.  Therefore, this script WILL produce two lines of code that you may or may not want.

ALTER DATABASE ‘ + DB_NAME() + ‘ SET COMPATIBILITY_LEVEL=100

and

ALTER DATABASE ‘ + DB_NAME() + ‘ COLLATE ‘ + @collation

where @collation is a variable defined at the beginning of the script.

You can alter the original script I’ve provided, or you can simply remove them from the output before saving.

The remaining results will be a collection of statements to alter the column collations:

ALTER TABLE [tbl] ALTER COLUMN [clm] column_spec() COLLATE [new_collation]

Once this has been executed, copy and paste the output into another script file for later use (myDatabase_post-script_2.sql)

Script 4 – Create Dependent Objects

This is where most of the work takes place.

You must also edit this script to use your desired collation.  Perform a search and replace of SQL_Latin1_General_CI_AS to your chosen collation.

Open the script, connect to the desired database and hit the [fire] button.  Again, like those previous, this script is non-destructive, and produces a series of DDL statements.

The first you’ll notice in the output is the “USE DBADB” statement.  It is very important to note here that if the stored procedure (sp_scriptObject) from the very first script is stored in another database, that this is reflected in the output.

The remainder of the output looks like:

EXEC DBADB.dbo.sp_scriptObject…

With all of the required parameters, based on the type and location of the object.  The other thing you may notice here is that the order in which these statements are in, is the exact reverse of the order from the 2^nd script; again, with the intent to avoid dependency crashes.

This output now needs to be copy and pasted to a new query window – not to a file.  The output of the new window should be set to text (Query > Results in Text) or (Ctrl-T).  Execute the statements.  Be aware, that this could take a significant amount of time based on the number of objects and how complex they are.

But, in essence, you will be trapping a buffer dump for each object that needs to be dropped.

VERY IMPORTANT.

There may be a case where an object’s DDL is simply too large.  There are limitations in which the VBScript engine stores string data.  I can’t recall exactly what the issue was, nor do I care to look it up again. But basically, the buffer is limited to 8k.  If your object DDL is greater than 8k, an error will be thrown and the object DDL will NOT be trapped.

After this script has finished, you must first scan the output for errors.  If any are found, identify the object which was not scripted and manually script those and paste them into the output.

This output must then be saved to a final script for execution against the target database.

Final Step

In the end, the above work will have produced 3 scripts that now need to be run against the target database.  Now is a good time to clarify that these generated scripts are very destructive.  So, please take your time, test, test and test.  If you choose to do this on a production box, your career could be at stake.  There is never one-size-fits-all solution, so for each scenario, these scripts may required modifications.  These were produced to fix my needs, but I did try to keep them as generic as possible for re-use.

On your target database, execute each of the generated scripts, obviously in the order specified.

The net result will leave your database with each column, along with the database set at your desired collation setting.

The Quick and Dirty Approach

This section is for those who don’t really care to know the details and simply provides an exact step-by-step.

1. Backup the target database.
2. Copy the 4 scripts to an empty folder.
3. Open Pre-Script_1_DROP_DEPENDENCIES.sql in a text editor and perform a search and replace of “Latin1_General_CI_AS” and change this to the desired collation.
4. Open Pre-Script_2_CHANGE_COLLATIONS.sql in a text editor and perform a search and replace of “Latin1_General_CI_AS” and change this to the desired collation.
5. Open Pre-Script_3_CREATE_DEPENDENCIES.sql in a text editor and perform a search and replace of “Latin1_General_CI_AS” and change this to the desired collation.
6. Create a database named “DBADB”
7. Execute Pre-Script_0_DBADB.sp_scriptObject.sql
8. Change the working database to the one which requires collation changes.
9. Execute Pre-Script_1_DROP_DEPENDENCIES.sql
10. Save the output to a text file named “dbName_post-script_1.sql”
11. Execute Pre-Script_2_CHANGE_COLLATIONS.sql
12. Save the output to a text file named “dbName_post-script_2.sql”
13. Execute Pre-Script_3_CREATE_DEPENDENCIES.sql
14. Copy the output to a new query window and execute.
15. Scan the secondary output for errors.  Identify any failed objects and manually script out the DDL.  Paste it into the location where the error occurred.
16. Now, save this output to a text file named “dbName_post-script_3.sql”
17. Copy dbName_post-script_1.sql, dbName_post-script_2.sql, and dbName_post-script_3.sql to the target machine.
18. Connect to the desired target database and execute each of the above post-scripts in numeric order.

Good luck.  Please heed any warnings I may have mentioned.  If you have any questions, certainly send me an email and I will try to answer promptly.  But, also please be patient, I do have a full time job that takes priority.

Brian Cidern
brians.sql.blog (at) gmail (dot) com

Reference : Pinal Dave (http://blog.SQLAuthority.com)

Many people assume that a change in the table structure is schema change. In fact, creating or dropping index on any table also qualifies as schema change.

As per the white paper, “Schema change” is defined as follows:

• Adding or dropping columns to a table or view.
• Adding or dropping constraints, defaults, or rules to/from a table.
• Adding an index to a table or an indexed view.
• Dropping an index defined on a table or an indexed view (only if the index is used by the query plan in question).
• Dropping a statistic (not creating or updating!) defined on a table will cause a correctness-related recompilation of any query plans that use that table. Such recompilations occur at that instant when the query plan in question begins execution. Updating a statistic (both manual and auto-update) will cause an optimality-related (data related) recompilation of any query plans that uses this statistic.

As per the above definition, I decided to experiment with schema change. In white paper, it was mentioned that schema change will recompile the execution batch. I decided to test this out. I created a small script (given below) and decided to test it.

As per the script, I created a large table and stored the procedure that was accessing that table. I created another object that was an index to the table; when SP was ran with index on table, it showed a plan different from that without the index on table. It was evident that the batch was recompiled when the index on the table was modified.

The script is available here to download.

Without Index execution plan:

With Index execution plan:

It is clear from the execution plan that it is different, and this proves the point that schema change recompiles the execution plan.

There is one more script provided in the white paper, which gets data from DMV related to cached plan. When I ran the following SP, it did not give me very interesting results.

SELECT cp.objtype AS PlanType,
OBJECT_NAME(st.objectid,st.dbid) AS ObjectName,
cp.refcounts AS ReferenceCounts,
cp.usecounts AS UseCounts,
st.TEXT AS SQLBatch,
qp.query_plan AS QueryPlan
FROM sys.dm_exec_cached_plans AS cp
CROSS APPLY sys.dm_exec_query_plan(cp.plan_handle) AS qp
CROSS APPLY sys.dm_exec_sql_text(cp.plan_handle) AS st;

When I ran the above script, I found that the UseCount column did not reset when QueryPlan is changed. This was very surprising as I was expecting UseCount to reset on changing the query plan. I sent an email to Dr. Greg Low, who is author of this white paper. He was very kind to respond back within minutes with a proper answer, which is as follows:

You’ll notice in Profiler when you run the sproc after creating the index that it does in fact cause a recompile. However, in SQL Server 2008 (and 2005), these are statement level recompiles not proc level recompiles. I imagine that’s why they’re not reflected in the UseCount values ie: the plan itself isn’t recompiled and that’s what the counts are based on.

Well, this simple explanation clarified my doubts. Let me know what you think, and I strongly suggest you all to read the white paper written by Dr. Greg Low.

Reference : Pinal Dave (http://blog.sqlauthority.com)

SQL SERVER – Discussion – Effect of Missing Identity on System – Real World Scenario

SQL SERVER – List All Missing Identity Values of Table in Database

Well what really interesting is that another user asked similar question to find gaps in sequence. Two different expert Brian and Tejas responded with excellent solution. Let us go over questions and their solutions.

Download script used in this article

Question by Enrico

Hi Pinal/Folks,

Can anyone please help me? I am stunned at this one.

Let’s say I have the following records:

INV0096
INV0097
INV0099
INV0100
INV0102
INV0103

How can I generate a SQL Script that will show me the gaps in the sequence?

In such a way that the results will give me INV0098 and INV0101.
Or even just the number 98, and 101.

Thank you so much for your time.

Best regards,

Enrico

First Solution by Brian Tkatch

/*
Brian Tkatch Solution 1
*/
WITH
Data(Datum)
AS
(
SELECT 'INV0096' UNION ALL
SELECT 'INV0097' UNION ALL
SELECT 'INV0099' UNION ALL
SELECT 'INV0100' UNION ALL
SELECT 'INV0102' UNION ALL
SELECT 'INV0103'
),
CTE
AS
(
SELECT
CAST(SUBSTRING(MIN(Datum), 4, 4) AS INT) Start,
CAST(SUBSTRING(MAX(Datum), 4, 4) AS INT) Finish
FROM
Data
UNION ALL
SELECT
Start + 1,
Finish
FROM
CTE
WHERE
Start < Finish
)
SELECT
Common.Formatted
FROM
CTE
CROSS APPLY(SELECT 'INV' + RIGHT('0000' + CAST(Start AS VARCHAR(4)), 4)) Common(Formatted)
WHERE
NOT EXISTS
(
SELECT
*
FROM
Data
WHERE
Data.Datum = Common.Formatted
)
OPTION
(MAXRECURSION 0);

Second Solution by Brian Tkatch

/*
Brian Tkatch Solution 2
The CROSS APPLY is just nice, but not required. Without it, the query is very similar:
*/
WITH
Data(Datum)
AS
(
SELECT 'INV0096' UNION ALL
SELECT 'INV0097' UNION ALL
SELECT 'INV0099' UNION ALL
SELECT 'INV0100' UNION ALL
SELECT 'INV0102' UNION ALL
SELECT 'INV0103'
),
CTE
AS
(
SELECT
CAST(SUBSTRING(MIN(Datum), 4, 4) AS INT) Start,
CAST(SUBSTRING(MAX(Datum), 4, 4) AS INT) Finish
FROM
Data
UNION ALL
SELECT
Start + 1,
Finish
FROM
CTE
WHERE
Start < Finish
)
SELECT
'INV' + RIGHT('0000' + CAST(Start AS VARCHAR(4)), 4)
FROM
CTE
WHERE
NOT EXISTS
(
SELECT
*
FROM
Data
WHERE
Data.Datum = 'INV' + RIGHT('0000' + CAST(Start AS VARCHAR(4)), 4)
)
OPTION
(MAXRECURSION 0);

Third Solution by Tejas Shah

/*
Tejas Shah Solution 3
*/
DECLARE @Test TABLE (Data VARCHAR(10))
INSERT INTO @test
SELECT 'INV0096'
UNION ALL
SELECT 'INV0097'
UNION ALL
SELECT 'INV0099'
UNION ALL
SELECT 'INV0100'
UNION ALL
SELECT 'INV0102'
UNION ALL
SELECT 'INV0103'
UNION ALL
SELECT 'INV0106'
UNION ALL
SELECT 'INV0110'
;WITH cte1 AS(
SELECT CAST(RIGHT(Data,4) AS INT) AS RowID
FROM @Test
), Missing AS(
SELECT MIN(RowID) AS MissNum,
MAX(RowID) AS MaxID
FROM Cte1
UNION ALL
SELECT MissNum + 1,
MaxID
FROM Missing
WHERE MissNum < MaxID
)
SELECT missnum
FROM Missing
LEFT JOIN cte1 tt ON tt.Rowid = Missing.MissNum
WHERE tt.Rowid IS NULL
OPTION (MAXRECURSION 0);


Let me know if you know any other trick to achieve the same.

Reference : Pinal Dave (http://blog.sqlauthority.com)

Off the top of my head, I can offer the following quick reasons for this sequence:

• The SQL Server Engine does not only run jobs for SQL Server Engine itself. It also runs jobs for other core services like the SQL Server Analysis Service, Integration Service, Reporting Service etc.;
• The SQL Server Agent can run almost any kind of task that an Operating system can run. For example invoking any program or running shell scripts;

• The SQL Server Agent also starts jobs which are scheduled to run the second the SQL Server Engine starts, and for this reason it is needed; and
• Replication, mirroring and a few other tasks also depend on Agent Jobs.

These are the reasons that I have come up with so far. Can you think of any more? Let us have your views.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

When providing the script, Imran asked me if I knew of any specific circumstances in which this kind of query could be useful, as he could not think of an instance where it would be necessary to find a missing identity. I was unable to think of a single reason for listing missing identities in a table. I posted Imran’s script on the assumption that someone would come up with an improved script, but as mentioned earlier, nobody did. Instead, we have been able to follow a very interesting discussion on subject of the need, if any, for listing Missing Identity values.

So, the question is this: “Do you know a real-world scenario where a Missing Identity value in any table can create problems?“

I have already received some extremely interesting comments from many experts, and all have posed the above question in one form or another. At this moment, I am still trying to think of an example from my own experience, but have yet to find one. Imran has since come up with one good example. Here is what he and other experts have suggested so far.

Jacob Sebastian – IDENTITY values are not expected to be sequential and there are all chances of having missing identity values, the most common cause is transaction rollbacks.

Simon Worth – The identity column is basically just a random number – even though they come sequentially. A developer making an assumption that the next record inserted will have an identity that is 1 more than the last inserted record. And if this is the case – then there are flaws in the logic of the developer.

Jacob Sebastian – What if the value in my table is 1, 2, 3, 5, 6 etc where “4″ is missing from the sequence. So what is the importance of knowing whether a table has missing identity values or not?

Imran Mohammed – If you use Identity property as your most unique column and Transaction Identifier, then definitely you would want to know why few transaction did not completely, Is there any specific fashion these transaction fails (Can be found out looking at missing values of identity)… Could be helpful to debug.

Now it is your turn. Let us have your thoughts.

Reference : Pinal Dave (http://blog.sqlauthority.com)

The best part about it is micro-blogging; you are not required to type a whole blog but just a statement of not more than 140 characters. Another advantage is that if you want to put a link then Twitter truncates the url to a tinyurl.com link, thus you can link a large url while saving space for typing more characters. Undoubtedly, Twitter is shrinking our world!

My twitter username is pinaldave.

There are already plenty of guides easily available about how twitter should be used and what are the advantages and disadvantages. You can find plenty of related information on the various sites, so we will not cover them today. In this post we will just see how I use it and what I exactly use it for.

I primarily use it for two reasons 1) to stay connected with my friends by reading their timeline 2) to post important news from my side or this blog on twitter. I use http://tweetdeck.com currently as my twitter client as it provides me multiple columns of timelines for all necessary groups. I can make my own personal group here as well.

Twitter Keywords

When I have to reply to someone I use @username for the person and that twit shows up on the person’s timeline as well as on my timeline. People who are following both of us will be able to view that reply as well. E.g. @pinaldave you love twitter.

When I have to inform everyone that someone has twitted me and I liked it, I simply prefix it with the word RT and that is broadcast to everyone following me. It will also show up as reply/mention of anybody’s name I mentioned in my message. E.g. RT @pinaldave you love twitter.

When I have to send direct email that can be viewed by only two people, I prefix it with the word D before my message to make it visible to those two; however, for this feature the other person has to follow you first. D pinaldave you love twitter.

I often use # for important searchable keywords. Many times, people search those hash tags using twitter search. E.g. Pinal Dave, you love #twitter.

Follow and Followers

I currently follow only people whom I genuinely want to follow. I do not follow everyone who comes across me. I think it is unfair to follow people if I am not interested in reading them. I keep my timeline open to let people follow me as they prefer and learn about exactly what is happening in my world.

A common adage we have often listened is – Quality is more important than Quantity. In Twitter this applies perfectly. It really does not matter to have a large number of people following you when they are not really reading your twits. When I follow someone, I give my 100% attention  and read everything they write.

Twitter is Personal Choice

I am not here to advocate the use of twitter. I would rather suggest that it is completely your choice. More importantly, overuse of twitter can for sure damage your online reputation.

I love twittering but when I am at work, I make sure that I do not leave the client open till I finish my work. It is not fair to my employer if I am twitting during my work time. I do not even respond to emails when I am at my consultancy work.

General Twitter Guidelines

There are numerous specialized guidelines for Twitter. Just treat this as your regular conversation. In regular conversation, we do not always repeat what other person suggests. Same way, one should control their RT (re-twitting). Also remember, no one will talk to us if we keep boasting about ourselves too much. Keep this in mind while twittering. I can go on writing about Twitter here, but I am sure you are pretty smart and can assume other rules by yourself.

The person from whom I learn tricks about twitter is Brent Ozar (@BrentO), I like his articles on twitter.

You are welcome to follow me on twitter here : pinaldave

Reference : Pinal Dave (http://blog.SQLAuthority.com)

My first plan is that I am going to take a break from blogging on anything technical today and spend more time with my family.

Let me tell you about my second plan. I am very much pleased and encouraged with active participation from many of you on this blog. I have received so far more than 10s of thousands of comments on my blog. These comments are open for you all to read and you can use them in any fashion that can help you out. Without good participation there is no life in any community, whether it is offline or online.

For the same reason, I have introduced Top 7 Commenters section in the side bar of SQLAuthority.com. There are many active users who have been participating for a long time. One of them I just recognized is Imran Mohammed. This SQL Expert has been extremely helpful to the community and has been providing comprehensive answers to hundreds of questions. Hats Off to you Imran! I totally appreciate your knowledge and dedication to SQL community. Without passion and love toward helping community, this is absolutely not possible.

You can subscribe to SQLAuthority.com feed, which will display all the new posts of my blog as they are published.

You can subscribe to SQLAuthority.com comments, which will show all the new comments posted on the blog as they are published.

If you prefer to receive SQLAuthority.com articles in email, you will get all the fresh posts in email.

There is no automated feature for top commenters where I host, so I will have to calculate this manually. I will keep it updated on weekly basis. Every week we will see new names there. First column is number of post, and second column is number name with link to experts website or blog or email.

Every day I receive a couple of emails about starting a forum on SQLAuthority.com. I do plan to start the forum in future. Currently, there are plenty of very good forums on the web. SQLAuthority.com is a blog and below every blog post there is a discussion related to the subject discussed in the blog post.

SQLAuthority.com is all about community and people who are passionate about SQL. Everyone is welcome to participate here. There are many ways to participate.

• Discussing blog post
• Helping user who have posted their issues on blog
• Asking questions
• Writing Guest Post
• Submitting useful SQL Scripts
• Reviewing any Microsoft-related products
• Reviewing Books

Let me know if you have any other useful and interesting suggestions on how I can involve YOU more in my efforts to take SQL community to the next level.

If you are interested in participating in volunteer program do send me email at pinal ‘at’ SQLAuthority.com and we can discuss something out.

Reference : Pinal Dave (http://blog.sqlauthority.com)

Let us take a look at an example where we will generate script database. In our example, we will just take one table for convenience.

Right Click on Database >> Tasks >> Generate Scripts >>

This will pop up Generate SQL Server Scripts Wizards >> Click on Next >> Select Database >> This will bring up a screen that will suggest to Choose Script Option.

On Choose Script Option Screen of Script Wizard under section Table/View Options Look at Script Data row and Turn the Option to True.

The Next Screen will ask about object types. Select all the objects that are required in the generated script. Depending on the previous screen it will show few more screen requesting details about the objects required in script generation.

On the very last screen it will request output options. Select the desired output options of Script to file, Script to Clipboard or Script New Query Window.3

Clicking on Finish button will generate a review screen containing the required objects along with script generating data.

Clicking on Finish button one more time will generate the requested output.

Similarly, if you want to generate data for only one table, you can right click on the table and follow almost a similar wizard. I am sure this neat feature will help everybody who has been requesting for it for a long time.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

The fact is that one can easily go to MSDN and learn the difference, so I head myself to MSDN and read the difference. This is what the Book On Line suggests here.

NORECOVERY specifies that roll back not occur. This allows roll forward to continue with the next statement in the sequence.
In this case, the restore sequence can restore other backups and roll them forward.
RECOVERY (the default) indicates that roll back should be performed after roll forward is completed for the current backup.
Recovering the database requires that the entire set of data being restored (the roll forward set) is consistent with the database. If the roll forward set has not been rolled forward far enough to be consistent with the database and RECOVERY is specified, the Database Engine issues an error.

Book On Line has very clearly explained the same subject and provides succinct explanation of the difference.

In other words -

While doing RESTORE Operation if you restoring database files, always use NORECOVER option as that will keep database in state where more backup file are restored. This will also keep database offline also to prevent any changes, which can create itegrity issues. Once all backup file is restored run RESTORE command with RECOVERY option to get database online and operational.

It is also important to be acquainted with the restore sequence of how full database backup is restored.

First, restore full database backup, differential database backup and all transactional log backups WITH NORECOVERY Option. After that, bring back database online using WITH RECOVERY option.

Following is the sample Restore Sequence

RESTORE DATABASE DATABASE FROM full_database_backup WITH NORECOVERY;
RESTORE DATABASE DATABASE FROM differential_backup WITH NORECOVERY;
RESTORE LOG DATABASE FROM log_backup WITH NORECOVERY;
-- Repeat this till you restore last log backup
RESTORE DATABASE DATABASE WITH RECOVERY;

I hope now it is very clear to you all what is restore sequence and the difference between recovery options.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

Although the Official time to commence the event was at 1:45 PM we were really excited to see the attendees entering the hall before the official time. We were more than happy to quickly arrange a special session for those attendees who came in early to keep them engaged. They won loads of gifts while waiting for the official sessions to start. We had a brainstorming quiz session. I was happy to see that people in Ahmedabad are now taking keen interest in enhancing their knowledge.

One of the questions was – What is the maximum number of Index per table? I received lots of answers to this question but only two answers are correct. Let us now take a look at them.

For SQL Server 2005:
1 Clustered Index + 249 Nonclustered Index = 250 Index
http://msdn.microsoft.com/en-us/library/ms143432(SQL.90).aspx

For SQL Server 2008:
1 Clustered Index + 999 Nonclustered Index = 1000 Index
http://msdn.microsoft.com/en-us/library/ms143432.aspx

I would like to thank one of the attendees of Ahmedabad TechEd Kruti Kansara for the above comment.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

SQL SERVER – Interesting Observation of Logon Trigger On All Servers

SQL SERVER – Interesting Observation of Logon Trigger On All Servers – Solution

If you are wondering what made me write yet another article on logon trigger then let me tell you the story behind it. One of my readers encountered a situation where he dropped the database created in the above two articles and he was unable to logon to the system after that.

Let us recreate the scenario first and attempt to solve the problem.

/* Create Audit Database */
CREATE DATABASE AuditDb
GO
USE AuditDb
GO
/* Create Audit Table */
CREATE TABLE ServerLogonHistory
(SystemUser VARCHAR(512),
DBUser VARCHAR(512),
SPID INT,
LogonTime DATETIME)
GO
/* Create Logon Trigger */
CREATE TRIGGER Tr_ServerLogon
ON ALL SERVER FOR LOGON
AS
BEGIN
INSERT INTO AuditDb.dbo.ServerLogonHistory
SELECT SYSTEM_USER,USER,@@SPID,GETDATE()
END
GO
/* Dropping Database AuditDB */
/* Please note login to Server again will
Produce Errors */
USE master
GO
DROP DATABASE AuditDB
GO

After the database is dropped and we try to login again the following error will be displayed.

Logon failed for login ‘SQL\Pinal’ due to trigger execution.
Changed database context to ‘master’.
Changed language setting to us_english. (Microsoft SQL Server, Error: 17892)

This error could have been evaded if the Trigger was dropped prior to dropping database.

USE master
GO
DROP TRIGGER Tr_ServerLogon ON ALL SERVER
GO

Now, it will not be possible to logon to the database using SQL Server Management Studio. The only way to fix this problem is using DAC. Read more details about DAC  Using a Dedicated Administrator Connection to Kill Currently Running Query. Now let us take a look at the example in our case where I am using windows authentication to logon to the system.

Connect SQL Server using sqlcmd and -A option, which will establish DAC. Running the following command on command prompt will allow you to login once again in the database. I am specifying argument -d master which will directly enable me to logon to master database when DAC is connected.

C:\Users\Pinal>sqlcmd -S LocalHost -d master -A
1> DROP TRIGGER Tr_ServerLogon ON ALL SERVER
2> GO

I hope my explanation on logon triggers, DAC, and sqlcmd is clear to everyone. Let me have your thoughts about my present article.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

You do not need to install all Service Packs to upgrade your server version to the latest versions. Just install Latest Service Pack and Hotfix and you are done!

This might be the smallest tip I have given so far on my blog but I am just curious to know from you all if you already knew about the solution or came to know about it through my blog today. Based on your feedback, I will come up with small articles like this in future and solve your day-to-day queries.

All Microsoft Service Packs are listed here.

Reference : Pinal Dave (http://blog.SQLAuthority.com)

To put the long story sort, let us today in this blog post try to understand what interim table is and how many possible interim tables are there.

Interim table is a table that is generated by joining two tables and not the final result table. In other words, when two tables are joined they create an interim table as resultset but the resultset is not final yet. It may be possible that more tables are about to join on the interim table, and more operations are still to be applied on that table (e.g. Order By, Having etc). Besides, it may be possible that there is no interim table; sometimes final table is what is generated when query is run.

Let us try to comprehend the above explanation with an example using the following AdventureWorks sample database.

USE AdventureWorks;
GO
SELECT *
FROM Person.Contact C
INNER JOIN Sales.SalesPerson SP
ON C.ContactID = SP.SalesPersonID
INNER JOIN Sales.SalesTerritory ST
ON ST.TerritoryID = SP.TerritoryID
ORDER BY ST.TerritoryID, C.LastName
GO

First Interim Table:

When two tables are joined and there are more pending operations then the table is called first interim table.

Second InterimTable:

When a third table is joined to the first interim table and there are more pending operations then the table is called second interim table.

Final Table:

After joins, all the remaining database operations are performed on the latest interim table which generates the final table.

Please let me know if your doubts about interim tables are clear now with this simple and easy-to-read post. If you like the style of this post, I will post many more like this in near future.

Reference : Pinal Dave (http://blog.sqlauthority.com)

“How can I find which object belongs to which filegroup. Is there any way to know this?”

Well, finding out which object belongs to which table is very simple.
Let us first create a database having multiple filegroups.

/* Create Database
Please note that there are two Filegroups for newly created datbase.
MDF - Primary and LDF on Secondary */
CREATE DATABASE [FGTest] ON  PRIMARY
( NAME = N'FGTest',
FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\FGTest.mdf' ,
SIZE = 3072KB , FILEGROWTH = 1024KB ),
FILEGROUP [Secondary]
( NAME = N'FGTest_2',
FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\FGTest_2.ndf' ,
SIZE = 3072KB , FILEGROWTH = 1024KB )
LOG ON
( NAME = N'FGTest_log',
FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\FGTest_log.ldf' ,
SIZE = 1024KB , FILEGROWTH = 10%)
GO

Now create a table on primary filegroup.

/* Create New table on Primary Filgroup */
USE [FGTest]
GO
CREATE TABLE [dbo].[TestTable](
[ID] [int] NOT NULL,
[Col1] [varchar](50) NOT NULL,
[Col2] [varchar](50) NOT NULL,
CONSTRAINT [PK_TestTable] PRIMARY KEY CLUSTERED
(
[ID] ASC
)
) ON [PRIMARY]
GO

Next, create non-clustered index on secondary filegroup.

/* Create Non Clustered Index on Secondary FileGroup */
CREATE NONCLUSTERED INDEX [IX_TestTable_Second] ON [dbo].[TestTable]
(
[Col1] ASC
) ON [Secondary]
GO

After that, we will run the following T-SQL and determine where all the objects are located on filegroup. We have already created a non-clustered index on our table. Now, non-clustered table creates its own table, which consists of pointers to clustered index. For the same reason, we will search with additional condition of object type, which will list only user created tables. Non- clustered table should be part of user created table group.

/* Get Details of Object on different filegroup
Finding User Created Tables*/
SELECT o.[name], o.[type], i.[name], i.[index_id], f.[name]
FROM sys.indexes i
INNER JOIN sys.filegroups f
ON i.data_space_id = f.data_space_id
INNER JOIN sys.all_objects o
ON i.[object_id] = o.[object_id]
WHERE i.data_space_id = f.data_space_id
AND o.type = 'U' -- User Created Tables
GO

Let us observe the same query again with different WHERE condition. Here we are retrieving everything that is on secondary filegroup. It is very easy to identify filegroup name and data by just selecting everything from system table sys.filegroups.

/* Get Detail about Filegroups */
SELECT *
FROM sys.filegroups
GO

In our case, secondary filegroup has data_space_id as 2. Now, we will run the following query and figure out which objects are located on filegroup 2.

/* Get Details of Object on different filegroup
Finding Objects on Specific Filegroup*/
SELECT o.[name], o.[type], i.[name], i.[index_id], f.[name]
FROM sys.indexes i
INNER JOIN sys.filegroups f
ON i.data_space_id = f.data_space_id
INNER JOIN sys.all_objects o
ON i.[object_id] = o.[object_id]
WHERE i.data_space_id = f.data_space_id
AND i.data_space_id = 2 -- Filegroup
GO

I hope this answers Jogi’s question and is helpful for those who were curious to know the answer to this topic.

Reference : Pinal Dave (http://blog.SQLAuthority.com)