Re: Embedded SQL weirdness

[Mark Lazarus <mlazarus@xxxxxxxx>]

Bruce,

 Thanks for your comments.

At 08:10 PM 1/18/99 -0500, you wrote:

>If SQL cannot find any rows that satisfy the search condition, an SQLCODE
of +100
>is returned."
>
>That last sentence begs the question on the padding issue again.  In the
other
>message, you also asked if this was different because it was imbedded.
You also
>state that you are on V3R2, so...

>First:  What cumm pkg are you at?  Not a cop out.  I have seen problems in
SQL
>that are addressed by later V3R2 cumm pkgs.  Check it out with a SNDPTFORD
>SF98320.  This will tell you where you *should*  be with cumm pkgs.

 The admin guys s/b applying the latest PTFs soon, but I think that we're
fairly current.  Also, they did a serach through the PTF database and
didn't come up w/ any relevant hits.

>Third:  Yeah, I am pretty darn sure about that padding thing.
Lexicographically,
>trailing blanks are usually not counted.  For example, say I have a database
>column that is 5 characters long.  Say I have a value in that column (I
will use
>lower case b for blanks and only uppercase to try an keep this clear) of
'AAbbb'
>and another row with the value of 'bAAbb'.  Now obviously, (or maybe not so
>obviously) if I look for the value 'AA', the first row hits, the second
row does
>not.  Why?  because the test case is padded with blanks to the length of the
>compared field, so the select predicate becomes 'AAbbb' itself.  So, did
you do a
>MOVEL to get the field built?  Is it really only 2 characters?  

 The 4 char field is coming from the DB, the 2 char is from the screen.

>Is the data in the table really left justified?  Does it really have two
spaces (true, hex '40'
>characters)?

 Yes and yes.

>Final suggestion?  Put the program in debug, run the program, look *very*
closely
>at the job log.  Make sure the program is not returning something in the
SQLCODE
>or SQLSTATE (the latter is more consistent as it is defined by the
standard while
>the former is implementor-defined).

 Sounds like a good idea.  Thanks again.

 -mark

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---