Tina,

Look at some of the fuzzy matching functions that SAS has, Soundex, Spedis, etc....

Toby Dunn

Hi,

I have a tough problem, at least for me, so I need and would really appreciate any help that you may have.

I want to pull some data from a data base for my dataset. I have a unique 8-digit company id (co_id) to do the matching.

After matching on 8 digits, first 7 digits, and first 6 digits of 'co_id', I still have ~2,000 firms unmatched out of ~9,000.

Next step, I want to match on 5 digits of 'co_id," and require that the matching firm have the first word of the sample firm's company name (coname_1stPart).

For example, if the sample firm is Walt Disney, I want to pull any firm

in the database that has the same 5 digits of 'co_id' and contain the word 'Walt' (coname_1stPart). I do not want to do exact matching (a.coname=b.coname), since the database may list the company name as 'Walt Disney Co.' or 'Disney (Walt)', etc.

I want to do something like "if index ( b.coname, 'a.coname_1stPart')." But the tricky part is I do not know 'a. coname_1stPart' before hand.

Any suggestions?

Thanks very much.

Tina