Any function can find out the similar numeric string?

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

shenanwei@gmail.com - 15 Jan 2008 16:21 GMT

I have 2 strings here which has 1 or 2 digits different because of
typo, like below
179705723 179705732

Any simple way I can make them similar.
soundex function is no use at all, it return every thing into Z000.

My idea is to compare them byte by byte, if only 2 bytes are
different, I will say it is same.
Thanks.

Reply to this Message

jefftyzzer - 15 Jan 2008 19:20 GMT

On Jan 15, 8:21 am, shenan...@gmail.com wrote:

> I have 2 strings here which has 1 or 2 digits different because of
> typo, like below
[quoted text clipped - 6 lines]
> different, I will say it is same.
> Thanks.

Sounds like you're looking for an edit-distance function, which
doesn't come with DB2 pre-built. Two well-known edit-distance
algorithms are Levenshtein distance and Smith-Waterman distance, and
it would not be difficult for you to build a DB2 UDF that implements
either one.

Have a look at these links:

http://en.wikipedia.org/wiki/Levenshtein_distance
http://en.wikipedia.org/wiki/Smith-Waterman_algorithm
http://www.merriampark.com/ld.htm

Regards,

--Jeff

Have a look at these sites

Reply to this Message