Re: Record by record update with SQL with no matching key.

Probably have to reorg the "live" member so that the rrns were sequential and in the same range, yes? (the file is quite old and hasn't been re-orged ever to my knowledge)

I may try that. I am on i 6.1

Pete


Dennis Lovelady wrote:

So now I would like to read through each
record
in the live data and replace 5 fields of information in it with data
from the fake file. Of course the fake file isn't keyed and wouldn't
have a matching employee number in any case, so the challenge is to
read
the live data record by record and replace it, record by record with
the
5 fields from the fake file. There are the same number of records in
both files.

I haven't worked this out completely, but you could join the two files even
though there is no common key. An example of this is as follows:

with t1 as (select columna cola, rrn(a) rrna from file1 a)
, t2 as (select columna colb, rrn(b) rrnb from file2 b)
select rrna, cola, colb from t1
join t2 on rrna = rrnb
(At least this works at V5R3. You didn't name your OS level.)

Does that help spawn any ideas?

Dennis Lovelady
http://www.linkedin.com/in/dennislovelady
--
"I always keep a supply of stimulant handy in case I see a snake -- which I
also keep handy."
-- W.C. Fields