Re: Newbie SQL problem with UPDATE - SQL0811 - repost

On Sun, Apr 12, 2009 at 2:35 PM, Tom Liotta <qsrvbas@xxxxxxxxxxxx> wrote:

Since I agree, I put it at more than 2 pennies. As I said, I'd probably
do it with SQL myself. I've used SQL for more than 20 years, though, and
the problem was clear, as it was to those who supplied the multiple answers.

What made me think was how a debug session would have shown exactly what
the problem was in RPG. In SQL, you either already know the answer or
you start asking to see if someone else knows. Debug isn't quite the
same with SQL. You don't get to see what happens inside the SQL
statement as it runs.

True as far as it goes, however it's usually pretty easy to pull out
parts of the statement and look at the intermediate results.

I often build up statements from simple to complex to make sure I'm
getting what I want to get.

Somehow, it seems like there should be a better way. As abstraction goes
to higher levels, things get harder to see.

I disagree, as I've found it easier to debug having the entire
intermediate set of records to look at instead of having to follow
each row all the way through the process one at a time.

Consider the OP, had he simply wrote the program in RPG, he probably
would not have even realized he had a problem! As he would have
needed to explicitly check for more than one record in LRICEXTWF by
using a chain + READE/READ instead of just chaining for the matching
record. I doubt he would have done this as it seems he was surprised
by the duplicates. The other thing that may save have saved the RPG
solution would have been if the OP created a unique keyed logical over
the exact keys he wanted.

Charles