EOFW not resume

["Gary Bowers" <gwbowers@xxxxxxxxxxxxx>]

I am a huge proponent of the trigger solution to detect when records have
changed in the database.

 

It has the advantage of happening right away (synchronous with the change),
and you don't have some polling program waiting for EOFDLY to see if there
might be something else to be done.

 

To be a bit more explicit:

 

ADDPFTRG FILE(PHYFILE)               

          TRGTIME(*AFTER)           

          TRGEVENT(*INSERT)         

          PGM(MYPROGRAM)  

 

The statement above would only be fired if record ADDs occur (new records
coming INTO the file)

 

ADDPFTRG FILE(PHYFILE)                 

         TRGTIME(*AFTER)               

         TRGEVENT(*UPDATE)             

         PGM(MYPROGRAM)                

         TRGUPDCND(*CHANGE)            

                                       

The statement above would only be fired if a record were updated AND the
record contents changed, which addresses your concern about someone changing
some other existing record to be an 'OS' record.

(If a record update occurred and all the fields were the same, the program
would not bother calling MYPROGRAM. - Yes, I have see plenty of programs
that still do this. )

 

 

There are plenty of good examples of trigger programs floating around, in
case you some. 

Check out the infocenter as well for more details about triggers.

 

 

 

PS: Depending on file activity, strongly consider not leaving the program
with *INLR on (keeps program active). 

[There, THAT should generate some comments]

 

Gary Bowers

Information Technology

Charlotte, NC