Discussion:
Sql 2005 Full-Text Web Searches with levenshtein and double metaphone
(too old to reply)
Ben Fidge
2007-04-24 08:27:26 UTC
Permalink
Hi,

Is it possible to incorporate levenshtein edit distance and double-metaphone
matching into Sql Server 2005's Full-Text Search queries using FREETEXTTABLE
or CONTAINSTABLE?

I'm working on a web-site that allows user to search multiple tables, and i
want the ability to pull back rows where the spelling is slightly out, or
contains words that sound like those provided in the query, a bit like
Google.

Also, I'd like to rank the results so that exact matches are returned at the
top of the result-set, and nearest-match or phonetic matches lower down the
list.

I've come across UDF's that provide these features but they don't appear to
be compatible with the FTS engine. I've also heard of alternative to the FTS
engine that integrate with Sql Server but I can't remember what they're
called.

Any suggestions?

Ben
Hilary Cotter
2007-04-24 10:47:52 UTC
Permalink
You can use the thesaurus option for some of this functionality, but you
have to know in advance the spelling differences and add them to the
substitute node of your thesaurus file.

I think the two alternatives you were thinking about were sql turbo and
lucene.
Post by Ben Fidge
Hi,
Is it possible to incorporate levenshtein edit distance and
double-metaphone matching into Sql Server 2005's Full-Text Search queries
using FREETEXTTABLE or CONTAINSTABLE?
I'm working on a web-site that allows user to search multiple tables, and
i want the ability to pull back rows where the spelling is slightly out,
or contains words that sound like those provided in the query, a bit like
Google.
Also, I'd like to rank the results so that exact matches are returned at
the top of the result-set, and nearest-match or phonetic matches lower
down the list.
I've come across UDF's that provide these features but they don't appear
to be compatible with the FTS engine. I've also heard of alternative to
the FTS engine that integrate with Sql Server but I can't remember what
they're called.
Any suggestions?
Ben
Ben Fidge
2007-04-24 11:15:36 UTC
Permalink
Hi Hillary

Will look into your suggestion, thank you.

Sql Turbo was indeed the one I was thinking about. Does Lucene integrate
into Sql Server, or is it just a stand-alone FT engine?

Thanks

Ben
Post by Hilary Cotter
You can use the thesaurus option for some of this functionality, but you
have to know in advance the spelling differences and add them to the
substitute node of your thesaurus file.
I think the two alternatives you were thinking about were sql turbo and
lucene.
Post by Ben Fidge
Hi,
Is it possible to incorporate levenshtein edit distance and
double-metaphone matching into Sql Server 2005's Full-Text Search queries
using FREETEXTTABLE or CONTAINSTABLE?
I'm working on a web-site that allows user to search multiple tables, and
i want the ability to pull back rows where the spelling is slightly out,
or contains words that sound like those provided in the query, a bit like
Google.
Also, I'd like to rank the results so that exact matches are returned at
the top of the result-set, and nearest-match or phonetic matches lower
down the list.
I've come across UDF's that provide these features but they don't appear
to be compatible with the FTS engine. I've also heard of alternative to
the FTS engine that integrate with Sql Server but I can't remember what
they're called.
Any suggestions?
Ben
Hilary Cotter
2007-04-24 11:48:11 UTC
Permalink
Its a standalone, however you can purchase one which is integrated into
SQL - dbsight - http://www.dbsight.net/
Post by Ben Fidge
Hi Hillary
Will look into your suggestion, thank you.
Sql Turbo was indeed the one I was thinking about. Does Lucene integrate
into Sql Server, or is it just a stand-alone FT engine?
Thanks
Ben
Post by Hilary Cotter
You can use the thesaurus option for some of this functionality, but you
have to know in advance the spelling differences and add them to the
substitute node of your thesaurus file.
I think the two alternatives you were thinking about were sql turbo and
lucene.
Post by Ben Fidge
Hi,
Is it possible to incorporate levenshtein edit distance and
double-metaphone matching into Sql Server 2005's Full-Text Search
queries using FREETEXTTABLE or CONTAINSTABLE?
I'm working on a web-site that allows user to search multiple tables,
and i want the ability to pull back rows where the spelling is slightly
out, or contains words that sound like those provided in the query, a
bit like Google.
Also, I'd like to rank the results so that exact matches are returned at
the top of the result-set, and nearest-match or phonetic matches lower
down the list.
I've come across UDF's that provide these features but they don't appear
to be compatible with the FTS engine. I've also heard of alternative to
the FTS engine that integrate with Sql Server but I can't remember what
they're called.
Any suggestions?
Ben
Loading...