RE: Data manipulation

Hi Tyra -

>>If it's always the last thing in the field it shouldn't
>>be too hard.  I would extract the last two sets of
>>non-blank characters in the field.
>
>Great idea....any idea on how to extract the last two sets
>of data?

I would use CHECKR to find the last non-blank character, then work to left
looking for a blank.  If I went more than 10 characters, then it is not a
valid U.S. or Canadian zip code, so I would skip the record.  Otherwise, I
would save what I had found in a field and repeat to get the next-to-last
set of non-blank characters.  If the next-to-last is more than 10
characters, I would not save it and just process the field with the last set
of non-blank characters.

Martin had a good idea with translating all letters to A and all digits to
1.  That will make pattern matching a LOT easier.  I would do that against
the extracted fields though, not the whole address field (unless the zip
code is not necessarily at the end).  After translation, if the last field
has ten non-blank characters and is 11111-1111, then it is persumably a U.S.
zip code with zip+4.  If it has five non-blank characters and is 11111, then
it is persumably a U.S. zip code without zip+4.

Ken
Southern Wine and Spirits of Nevada, Inc.
Opinions expressed are my own and do not necessarily represent the views of
my employer or anyone in their right mind.