what is fastest way to search text field with 'like'

we have 2 million rows with a comoany name column the problem is that there maybe different spellings for the same company name, i.e. jewel, jewels, jewel inc., etc.

indexing doesn't help, and like does full scan. any ideas to deal with issues like this? maybe some way of decoding or something? for example creating a field that has first 2 letters of name and then indexing that?

any ideas appreciated Received on Sun Jun 03 2001 - 15:05:08 CDT