Re: SQL update vs RPG/DDS update

Joe,

I see that I should have added more documentation with the code.

savmstr DS is the file from the first chain when the user entered the
program.
newmstr DS is the changes made in the program, but not yet sent to the file
mstrDS DS is reset at the last chain to whatever the file currently contains

mstrDS is the only DS updated when the file is read or chained to.
The others are updated through savmstr = mstrDS and newmstr = mstrDS
statements.


// if changes have been made, check file for changes
if newmstr <> savmstr;

// get the current record from the file
chain (cmyrin: cmycin) clmmas; // loads mstrDS

// compare the current record in the file to the original before
// making changes
if savmstr = mstrDS; If the file was changed by another
user these will not be equal
mstrDS = newmstr; // set fields to new values
clchgd = today;
clchgt = %time();
clchgu = userid;
update clmmas;
savmstr = mstrDS; // update saved DS to new values
origscreen = screenDS;
completed = *on;
else; // file has been changed by another user
unlock hclmmas; The lock is then released, however
clchgu does not show the update from the file
completed = *off;
message = NoUpdMst + %trim(clchgu) + Retry;
wrkadj = *zero;
wrkadj2 = *zero;
wrkadj3 = *zero;
endif;
else;
// no changes have been made, do not update
completed = *on;
//message = *blanks;
wrkadj = *zero;
wrkadj2 = *zero;
wrkadj3 = *zero;

// get the current record from the file
chain(n) (cmyrin: cmycin) clmmas; // loads mstrDS

Deletes are not allowed on this file, so there was no need to check the
found condition. If you got this far the record is known to exist.
Hopefully, this explains why I am puzzled. If I am just not understanding
what is going on here with these DS please enlighten me, because this is
driving me nuts!

Thanks for your time.

Gene

On 10/5/07, Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx> wrote:

From: Gene Burns

CHAIN followed by UPDATE, but with an externally defined DS.

fhclmmas uf E K DISK
| d mstrDS e ds extname(hclmmas)


// get the current record from the file
chain (cmyrin: cmycin) clmmas; // loads mstrDS

// compare the current record in the file to the original

before

// making changes
if savmstr = mstrDS;
mstrDS = newmstr; // set fields to new values
clchgd = today;
clchgt = %time();
clchgu = userid;
update clmmas;
savmstr = mstrDS; // update saved DS to new values
origscreen = screenDS;
completed = *on;
else; // file has been changed by another user
unlock hclmmas;
completed = *off;
message = NoUpdMst + %trim(clchgu) + Retry;
wrkadj = *zero;
wrkadj2 = *zero;
wrkadj3 = *zero;
endif;
else;


When the file has been updated by someone else the DS are different,
HCLMMAS
is unlocked, and the message is (supposedly) set telling the user who
updated the file.

I'm not sure what you're saying. The code is fine, and you should get
errors if someone changes the record (provided savmstr and newmstr are the
correct size).

The CLCHGU field is in the file, but is blank at this
point in the program. Checking the file shows that these fields have

been

updated and CLCHGU is not blank.

At WHAT point? Since the "mstrDS = newmstr" wipes out the original data,
the only place where the fields are guaranteed to match the file is the if
statement immediately after the chain but before the mstrDS = newmstr is
executed. If at that point you see the file different than what you have
in
your program, then you have a problem I have never encountered in my
career.

Anything other than that is more than likely a logic problem.

Joe

P.S. You also aren't checking for a not found condition on the chain, and
you definitely should in case someone deleted the record.

--
This is the RPG programming on the AS400 / iSeries (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/rpg400-l.