Practical Uses of PatIndex

  • Comments posted to this topic are about the content posted at http://www.sqlservercentral.com/columnists/rdavis/practicalusesofpatindex.asp


    My blog: SQL Soldier[/url]
    SQL Server Best Practices:
    SQL Server Best Practices
    Twitter: @SQLSoldier
    My book: Pro SQL Server 2008 Mirroring[/url]
    Microsoft Certified Master: SQL Server, Data Platform MVP
    Database Engineer at BlueMountain Capital Management[/url]

  • I just need now to find a need to use this PatIndex tool but it's good to now it exists

    Very interresting !

  • I Wonder if there if there is any difference in performance?  Anyone try any testing?

     

    Mark

  • Difference between using PatIndex and CharIndex, you mean? Good question. I have not done any such testing, but it might make for a good follow up article.


    My blog: SQL Soldier[/url]
    SQL Server Best Practices:
    SQL Server Best Practices
    Twitter: @SQLSoldier
    My book: Pro SQL Server 2008 Mirroring[/url]
    Microsoft Certified Master: SQL Server, Data Platform MVP
    Database Engineer at BlueMountain Capital Management[/url]

  • Don't see any need for the "IF" at the start of the function.  I wrote an almost identical function for this purpose but with just the WHILE.

    SQL DBA,SQL Server MVP(07, 08, 09) A socialist is someone who will give you the shirt off *someone else's* back.

  • One of the biggest differences is in application to the text fields. The charindex is not working (only for first 8000) while the patindex is. It returns the position while the like operation doesn't.

  • It is a very good post. However, it does not check for the valid range of integer number. For example: 3147483647 will result in true, however, according to sql server, this is not a valid integer because it exceeds the range of integers. ie -2,147,483,648 through 2,147,483,647.

    It also fails to capture comma separated integers, eg 1,234. This is a valid integer, but the UDF returns false.

    thanks,

    Mashiku

  • Thanks for the comments!! The UDF wasn't intended to be an all encompassing way to determine if a number could be converted to int. The UDF was merely a way to demonstrate how to use PatIndex effectively.


    My blog: SQL Soldier[/url]
    SQL Server Best Practices:
    SQL Server Best Practices
    Twitter: @SQLSoldier
    My book: Pro SQL Server 2008 Mirroring[/url]
    Microsoft Certified Master: SQL Server, Data Platform MVP
    Database Engineer at BlueMountain Capital Management[/url]

  • Here is a way that I recently found using PATINDEX useful:

    http://codeslammer.wordpress.com/2008/09/02/sql-server-string-manipulation-removing-non-numeric-characters/

    I have a column that contains an ID number, but there may be text before and/or after it. The text is often similar, but it can not be guaranteed that there will be text or not. However there should only be one number and this is the number I am trying to extract.

    For example, if you had a column with values such as:

    "Serial Number: 3078203984-02"

    "SN#472038572289"

    "Serial No. 02-28349-2074"

    and you just wanted to strip out the unique serial numbers, this is one option to do it.

  • In your article you say:

    Since this article is focusing on practical uses of PatIndex, I will only be checking to see if the string value contains digits only with the exception of allowing it to begin with a negative sign. I will not include checking to see if the value is within the allowable range of the Int data type.

    But even in your test:

    Select IsNumeric('-4'), PatIndex('%[^0-9]%', '-4'), dbo.fnIsInt('-4')

    It returned 0, so it wasn't doing what you said it was supposed to. (It isn't as simple as just two PatIndex calls.)

Viewing 10 posts - 1 through 9 (of 9 total)

You must be logged in to reply to this topic. Login to reply