Please start any new threads on our new site at https://forums.sqlteam.com. We've got lots of great SQL Server experts to answer whatever question you can come up with.

 All Forums
 SQL Server 2008 Forums
 Other SQL Server 2008 Topics
 Fuzzy searching

Author  Topic 

GEW1
Starting Member

3 Posts

Posted - 2013-07-24 : 08:53:28
Hello everyone,

I'm having trouble finding a way to use fuzzy searching (specifically Levenstein) in my database.

For now, I've been using an optimized T-SQL implementation of the Levenstein algorithm that I found on another forum. However, it is way too slow for me to use.

I'm unable to use CLR functions on my server due to "memory pressure" issues (My dba is nervous about using the -g start up parameter to allocate additional memory).

In addition, I'm unable to use Master Data Services' Similarity function. I'm using SQL 2008 R2 Standard, which doesn't support MDS.

Can anyone think of another way that I can use fuzzy searching to compare entries in my database?

Thanks in advance.

SwePeso
Patron Saint of Lost Yaks

30421 Posts

Posted - 2013-07-24 : 14:02:28
See http://levenshtein.blogspot.se/



Microsoft SQL Server MVP, MCT, MCSE, MCSA, MCP, MCITP, MCTS, MCDBA
Go to Top of Page

GEW1
Starting Member

3 Posts

Posted - 2013-07-25 : 13:41:15
Thanks for the reply but, as stated in my first post, I'm not working in an environment that lets me use CLR functions.
Go to Top of Page

SwePeso
Patron Saint of Lost Yaks

30421 Posts

Posted - 2013-07-25 : 17:51:35
You can implement a function, but performance will suffer.
See http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=51540



Microsoft SQL Server MVP, MCT, MCSE, MCSA, MCP, MCITP, MCTS, MCDBA
Go to Top of Page

GEW1
Starting Member

3 Posts

Posted - 2013-07-26 : 09:05:19
quote:
Originally posted by SwePeso

You can implement a function, but performance will suffer.
See http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=51540


I've been implementing different functions (including this one). It looks like any SQL implementation is going to be too slow due to how many rows I'm comparing. Thanks for the suggestions.
Go to Top of Page

SwePeso
Patron Saint of Lost Yaks

30421 Posts

Posted - 2013-07-26 : 14:16:05
The number of rows does affect computation speed. However, SQL Server have never been strong when working with strings.
And as the algorithm is written, it is a multiline scalar function which is even worse for performance in SQL Server.



Microsoft SQL Server MVP, MCT, MCSE, MCSA, MCP, MCITP, MCTS, MCDBA
Go to Top of Page
   

- Advertisement -