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.
| Author |
Topic |
|
phixx
Starting Member
1 Post |
 Posted - 2004-11-09 : 11:59:06
|
| Hello-I've recently been given the task of seeking out a solution to my firms Full-Text searching woes. My company is currently running MS-SQL 2000 Enterprise on a Windows 2003 machine. We have a database composed of full company names - around a million and a half records total. We have a functional full-text index in place, but right now searching for company names can be a bit tricky. Sometimes it fails because of weird crap that is in the company's name or alias.Because FTI only searches the beginnings of words, such as being able to return IBM by searching for "Int Bus Mac" but not by "nternational usiness achines", it has difficulties when the beginning has some odd punctuation. For example, something like "i-Link" is notoriously hard to search for.I need to determine the rules to use for this standardization algorithm. Some things are pretty obvious (take out all punctuation, etc.) but I need a comprehensive list of recommendations. Any ideas? I'm totally lost on this one |
|
|
AndrewMurphy
Master Smack Fu Yak Hacker
2916 Posts |
Posted - 2004-11-10 : 07:34:06
|
| Clean the input data is the best solution. You could also put in a workflow type solution, whereby new customer names are submitted for approval (by more experienced colleagues (and/or expert system))/qa'd for adherance to your desired standards before being accepted/modified within the database. |
 |
|
|
|
|
|
SQL Novice - Need Advice for Full Text Indexing