Re: How to extract single characters in PL/SQL?

noergl_at_gmx.net (Ludwig Nörg) writes:

> thanx a lot for your postings - but 2 hours after my first posting I
> found that solution you described by myself in the horrible Oracle 9
> Online SQL Reference. So the instr() helps, but I am in the terrible
> situation to split about 40 million customer data sets (input since
> 1973 ...) in totally 15 different occurrences :-(
>
> So it seems there should be a lot of manual post-editing if it doesn't
> work ...

Yuk! This is not going to be pleasant!

It might be worth while doing a few tests on your data sets and possibly developing some heuristics which might help reduce the data set which will need to be handled differently/individually. Perhaps extracting those fields which have more than 2 "words" for example might help identify people with names like Van Dyk etc.

I hope your 40M data set doesn't have lots of names from Asia, Middle East etc as this could make it even more difficult as there is a lot of variance in name structures and what is considered to be the family name etc.

good luck,

Tim

-- 
Tim X.
tcross (at) northnet com au