How to change word-break characters in SQL Server Full-Text indexing

Question

By default, when one tells SQL Server (currently using 2008) to Full-Text index a column, it treats characters such as \"@\" and \".\" as work-breakers, similarly to \" \".<

Answer 1

According to TechNet's article on SQL 2008 Full-Text Search:

well-known published interfaces provide the framework for Full-Text Engine extensibility. For more information, see the Microsoft Developer Network (MSDN) topics IFilter, IWordBreaker, and IStemmer.

So, at least according to this article, you can implement a custom IWordBreaker implementation (see http://www.siao2.com/2005/03/14/395199.aspx for more info) and get SQL to use it.

What I haven't found so far is how to plug your custom word-breaker into SQL itself-- how to tell SQL to use your word-breaker. Sorry for the incomplete answer... hope I got you at least part of the way to a solution.

Answer 2

In order to make your word breaker fly with SQL Server you have to disable signature verification and add your COM CLSID to the registry. For more info check out this post: http://blogs.msdn.com/shajan/default.aspx It helped me a lot! However I never managed to create my own language so I simply hijacked an already existing one.