Using REPLACE in an UPDATE statement

By Bill Graziano on 31 March 2010 | Tags: UPDATE , Functions

This article covers using the REPLACE function to selectively replace text inside a string in SQL Server. The REPLACE function is easy to use and very handy with an UPDATE statment.

Replace searches for certain characters in a string and replaces them with other characters. So this statement:

SELECT Replace('SQLTeam.com Rocks!', 'Rocks', 'Rolls')

will return

SQLTeam.com Rolls!

REPLACE searches the the first string for any occurance of the the second string and replaces it with the third string. You can also do replacements of different sizes. For example,

SELECT Replace('SQLTeam.com Rocks!', 'Rocks', 'is cool')

gives us

SQLTeam.com is cool!

I replaced a five character string with a seven character string with no problem. If the string isn't found, no changes will be made.

SELECT Replace('SQLTeam.com Rocks!', 'Yak', 'Tibetan bison')

returns exactly what we started with which is

SQLTeam.com Rocks!

If it doesn't find anything to change it just returns the string unchanged. You can use REPLACE in an UPDATE statement. Using the pubs database we could write:

Update dbo.authors
Set    city = replace(city, 'Salt', 'Olympic');

There were two authors that had "Salt Lake City" in the CITY field. Now that field holds "Olympic Lake City" for those two authors. The CITY field is unchanged for all the other authors.

A more common approach is to use this in conjuntion with a WHERE clause like this:

UPDATE dbo.authors
SET    city = replace(city, 'Salt', 'Olympic')
WHERE  city LIKE 'Salt%';

This only affects the rows that start with 'Salt'.

Note: This article was originally published in January 2002 and has been updated through SQL Server 2008 R2.