Generating Scripts only for Clustered and Non-Clustered indexes in SQL Server 2005

vishah — Thu, 05 Apr 2007 23:31:00 GMT

In following cases you will need to generate a scripts for all the clustered and non-clustered indexes on the SQL Server 2005 database:

1. During heavy bulk inserts into a table, you might want to drop all the indexes on the table to make the bulk load process go faster. Once the bulk load is done you can re-create those indexes back using the scripts.

2. If all the clustered and non-clustered indexes are built on a one single filegroup (PRIMARY, for example) and if you would like to move them to an appropriate filegroup (_DATA and _IDX, for example), you would need both create and drop scripts for clustered and non-clustered indexes.

There is a way to achieve that. You can download a utility called DBScript from below website:

http://www.codeplex.com/SQLSrvScrpt/Wiki/View.aspx?title=DBScript%20Utility&referringTitle=Home

This utility will help you script clustered, non-clustered indexes, foreign keys, and triggers - both CREATE and DROP.

Indexing a SQL Table

vishah — Thu, 01 Mar 2007 20:21:00 GMT

I found one of my old documents where I kind of summarized Indexes on a SQL Table, might help.

Indexing SQL Table

Primary Key:

· Creates clustered index on the column

· Doesn't allow NULLS

Unique Key:

· Creates non-clustered index on the column

· Allows only one NULL

Index on a table in general:

· Good for:

Ø Columns that contains large number of distinct values. If there are very few distinct values such as 0 and 1, most queries will not use index because in this case table scan is more efficient

Clustered Index:

· Good for:

Ø Columns that returns a range of values using operators such as BETWEEN, >, >=, <, and <=

Ø Columns that are accessed sequentially

Ø Columns that are frequently accessed by queries involving join or GROUP BY clauses

Ø Environment where very fast single row lookup is required. Create a clustered index on Primary Key column on such environment

· Not Good for:

Ø Columns that undergo frequent changes - High volume transactional environment where data tends to be volatile

· Benefit:

Ø If you apply cluster index on the columns which is used by ORDER BY and GROUP BY clause, that will eliminate SQL Server to sort data because the rows are already sorted. That will enhance the performance.

Non-Clustered Index:

· Good For:

Ø Columns frequently involved in search conditions of query (WHERE clause) that returns exact matches.

Ø Applications where joins and grouping are frequently involved. Create multiple non-clustered indexes on columns involved in join and grouping operations, and a clustered index on any foreign key columns.

Vipul Shah's SQL Blogs : Indexes

Generating Scripts only for Clustered and Non-Clustered indexes in SQL Server 2005

Indexing a SQL Table