Quantcast

SQL Server Indexes

Get the WebProNews Newsletter:
[ Business]

A database index is similar to an index in a book – it is comprised of a lookup value, and a number identifier that corresponds to the row number in a table.

In SQL Server, there are two kinds of indexes – clustered and non-clustered. Clustered Indexes require that the data in the table is physically sorted in the order of the index.

Because the data in a table can be physically sorted only one way, there can be at most only one clustered index per table.

Non clustered indexes do not require that data to be physically sorted, so there can be more that one non-clustered index per table.

In fact SQL Server allows up to 249 non-clustered indexes per table. Because data is not physically sorted, range searches using a non clustered index are not very efficient.

The command for creating an index in T-SQL is

CREATE [ UNIQUE ] [ CLUSTERED | NONCLUSTERED ] INDEX index_name ON { table | view } ( column [ ASC | DESC ] [ ,...n ] ) [ WITH < index_option > [ ,...n] ] [ ON filegroup ] < index_option > :: = { PAD_INDEX | FILLFACTOR = fillfactor | IGNORE_DUP_KEY | DROP_EXISTING | STATISTICS_NORECOMPUTE | SORT_IN_TEMPDB }

PAD_INDEX specifies the percentage of space left free on the non-leaf levels of the index.

FILLFACTOR specifies the percentage to fill the leaf pages.

SORT_IN_TEMPDB specifies that intermediate results of the sort will be stored in tempdb. This increases disk space requirement but affects speed index creation.

STATISTICS_NO_RECOMPUTE tells the system not to automatically update index statistics.

Of course, indexes can also be created and managed using the Enterprise Manager. They can be created using the Create Index Wizard, from the Database Diagram, or by modifying fields in the Table Designer.

There is a trade off with indexes. While they speed up execution of queries immensely, there is overhead associated with them. They consume additional disk space, and require additional time to update themselves whenever data is updated or appended to a table.

When loading large amounts of data it may pay to drop the index prior to the loading, then recreate the index after the new records have been appended to the table. Indexes can be dropped using the Table Designed, or by using the DROP INDEX command.

Indexes can also become fragmented. To defrag an index, either drop and recreate the index, or issue the command dbcc indexdefrag.

Add to | Digg | Yahoo! My Web

Technorati:

Metro NY / NJ SQL Server Consultants
We specialize is custom database software. Call us for a free consultation (973) 635 0080 or email us at paladn.com

Chris Kemp is a well known author in the field of Information Technology. His articles are very popular and well known in the various article banks across internet. His popular articles are about SQL Server, Database Design, IT Consulting and Software Development.

chriskemp@webpopularity.org

SQL Server Indexes
Comments Off
About Chris Kemp
Metro NY / NJ SQL Server Consultants
We specialize is custom database software. Call us for a free consultation (973) 635 0080 or email us at paladn.com

Chris Kemp is a well known author in the field of Information Technology. His articles are very popular and well known in the various article banks across internet. His popular articles are about SQL Server, Database Design, IT Consulting and Software Development.

chriskemp@webpopularity.org WebProNews Writer
Top Rated White Papers and Resources

Comments are closed.