Luke Hayler - SQL Server Developer » T-SQL

T-SQL Script to Identify Space Used By Each Table in a Database

Luke Hayler — Mon, 12 Sep 2011 21:21:16 +0000

There are often times that I need to do checks on the space used by databases on the server, and in particular, the tables in those databases. This is a script that I quickly put together that will return the following information for each of the tables in the database on which the script is run. The following information is returned:

Table Name
Row Count
Reserved Space (KB)
Data space used (KB)
Index size (KB)
Unused space (KB)

The above information is gathered using sp_spaceused. Now I’m sure that there are a number of other ways to get this information, but this was quick and easy to produce and use. If you have anything that you’d like to contribute, I’d love to hear it.

/***
    Script to identify the space used by each table in a database
    Information returned includes:
        - Table name
        - Record Count
        - Reserved Space (KB)
        - Data Size (KB)
        - Index Size (KB)
        - Unused Space (KB)
***/
 
/* Create a table variable to hold the values returned by sp_spaceused  */
DECLARE @Results TABLE 
(
      RowNum            INT IDENTITY(1,1)
    , TableName         VARCHAR(200)
    , Records           INT
    , ReservedSpace     VARCHAR(100)
    , DataSpace         VARCHAR(100)
    , IndexSizeSpace    VARCHAR(100)
    , UnUsedSpace       VARCHAR(100)
    , CheckDateTime     DateTime DEFAULT GETDATE()
)
 
/* Declare a variable to hold the Object Names */
DECLARE @ObjectName VARCHAR(200)
 
/* 
    Declare a cursor to loop through each of the tables in the database 
    and return their space used information. 
*/
DECLARE TableCursor CURSOR
FOR
    SELECT 
        ObjectName = TABLE_SCHEMA + '.' + TABLE_NAME
    FROM 
        INFORMATION_SCHEMA.TABLES
    WHERE
        TABLE_TYPE = 'BASE TABLE'
    
OPEN TableCursor
 
FETCH NEXT FROM TableCursor
INTO @ObjectName
 
WHILE @@FETCH_STATUS = 0
BEGIN
 
    /* Execute sp_Spaceused for each table and load the results into the table variable  */
    INSERT INTO @Results 
    (
          TableName            
        , Records            
        , ReservedSpace        
        , DataSpace            
        , IndexSizeSpace    
        , UnUsedSpace
    )
 
    /* 
        Uncomment the @updateusage option if you want to ensure your stats are up-to-date
        However, this may take some time depending on the size of your database tables. 
        It is set to 'false' in sp_spacedused by default.
    */    
    EXEC sp_spaceused @ObjectName
    --, @updateusage  = 'true'        
    ;    
    
    FETCH NEXT FROM TableCursor
    INTO @ObjectName
END
 
CLOSE TableCursor
DEALLOCATE TableCursor
 
/*    
    Get the list of tables and their associated space usage data.
    Update to your preferred format
*/
SELECT
      TableName            
    , Records            
    , ReservedSpace        = CAST(SUBSTRING(ReservedSpace, 1, LEN(ReservedSpace)-3) AS INT)
    , DataSpace            = CAST(SUBSTRING(DataSpace, 1, LEN(DataSpace)-3) AS INT)
    , IndexSizeSpace    = CAST(SUBSTRING(IndexSizeSpace, 1, LEN(IndexSizeSpace)-3) AS INT)
    , UnUsedSpace        = CAST(SUBSTRING(UnUsedSpace, 1, LEN(UnUsedSpace)-3) AS INT)
FROM @Results
ORDER BY 
    TableName

Copy. Paste. Enjoy.

**UPDATE – 2011/09/14**

My friend Jim McLeod (Blog | Twitter) has informed me that there is a better way to get the space used and row count details for objects in your database. His script utilises the sysindexes table, which means that this is compatible across a wider set of SQL Server versions including 2000/2005/2008. His original script appears in the comments below. I have reproduced it here with a little formatting, commenting and an extra table to restrict the results to just *user* tables:

/***
    This script will generate a list of User tables
    and their associated space used and row count details.
***/
 
SELECT 
      SchemaName    = ss.name
    , ObjectName    = OBJECT_NAME(si.id)
    , ReservedMB    = reserved * 8/1024.0
    , DataMB        = dpages * 8/1024.0
    , RecCount      = rowcnt   
 
FROM 
            sysindexes si
INNER JOIN  sys.objects so  ON    si.id = so.object_id
INNER JOIN    sys.schemas ss    ON    so.schema_id = ss.schema_id
WHERE
/*    Limit the resultset to bring back Heaps & Clustered indexes only 
    (Heap = 0, Clustered Index = 1)
*/
    indid < 2        
/*    We only want to see the 'user' tables at this stage. 
    You can include System tables by adding 'S' to the xtype clause 
*/
AND type = 'U'        
ORDER BY 
    SchemaName,
    ObjectName

I’ve left the first script in the post for comparison. Thanks Jim!

Refresh Intellisense in SQL Server Management Studio 2008

Luke Hayler — Tue, 01 Mar 2011 04:00:00 +0000

SQL Server Management Studio 2008 has been around for a little while now, but I still hear loads of people complaining that intellisense does not update often enough (or at all in some cases). This is pretty frustrating. I should know, I battled with this until I found the Intellisense menu and the shortcut keys.

Go to the SSMS Menu and click on Edit:

Now click on Intellisense> to view the Intellisense options

The Refresh Local Cache will do exactly that and then you’ll be able to see those changes that you have just implemented.

The shortcut (if you are a little visually impaired) is Ctrl + Shift + R

Now go refresh your cache.

Set Parameter Defaults for Stored Procedures

Luke Hayler — Mon, 17 Jan 2011 04:00:00 +0000

Did you know that you can set defaults for your Stored Procedure parameters? No? Well, it turns out that you can. And it’s pretty straight forward too.

To set a default to a stored procedure parameter all you have to do is add “= ” when you declare your parameters in your CREATE PROCEDURE statement.

CREATE PROCEDURE usp_MyStoredProcedure (@Parameter1 DATE = GetDate(), @Parameter2 VARCHAR(10) = ‘MyValue’)

That’s it!

Rediscovering RANK() – Selecting the TOP N Rows for each Group or Category

Luke Hayler — Tue, 06 Apr 2010 09:00:00 +0000

There may be times when you need to get the top 10 (or n) rows from each group of your dataset to satisfy a business requirement. There are a few ways this can be done, but none so elegant as when you use RANK().

RANK() effectively applies a ranking to each rows in your dataset, based on a specified grouping and order. This can then be used to select the Top n rows by using an appropriate where clause. Here’s an example:

   1: USE Adventureworks

   2: go

3:

   4: WITH OrdersCTE

   5: AS

6:

   7: (

   8: Select

   9:     salespersonid,

  10:     customerid,

  11:     SUM(subtotal) Total,

  12:     RANK() OVER(PARTITION BY salespersonid ORDER BY SUM(subtotal) desc) Ranking

  13: from Sales.SalesOrderHeader

  14: Where

  15:     salespersonid is not null

  16: Group by

  17:     customerid,

  18:     salespersonid

  19: )

20:

  21: Select

  22:     salespersonid,

  23:     customerid,

  24:     Total,

  25:     Ranking

  26: From OrdersCTE

  27: Where

  28:     Ranking <= 10

In the above example, I needed to find the top 10 Customers for each Sales Person based on the sum of all order subtotals (using AdventureWorks 2005). Here are the results:

Ranking Results

You can read more about the RANK() function here, and for more ranking functions see DENSE_RANK(), NTILE() & ROWNUMBER().