That's really strange. I'm working in RPG in AS400 for many years and I had never heard anything like this. Are you <b>absolutely sure</b> that the program you are running is the one corresponding to the RPG source? Do you have compiled it yourself? Perhaps you can check also the job stack while in the breakpoint and check the name and library (with SYSREQ, ...). Perhaps can exist another pgm with the same name performing a delete instead of a write... (I'm just guessing!). Check for pgm's with that name in all libraries (WRKOBJ *ALL/pgmname). Are you really sure that you do not have a trigger associate with that file? Are you journaling the file? If you are, you can display the "delete" journal entry and see what is the name of the program that is performing the write and the delete (take note: BOTH ENTRIES SHOULD BE IN THE JOURNAL). If do not have a journal, you can create one for the only purpose of finding what is going one with this issue. Do you have any OVRDBF? Do you use exit programs? I'll be listening for news in this. <b>Ok - here's the deal: I found another job running in a different subsystem that was picking up records and deleting them. It shouldn't have had that library in its list, but did. I'm still amazed at how quickly the update happened though because the other program was on a 5 second delay between I/O. I should have seen a record in the file before it was deleted, but never did.</b>

That's really strange. I'm working in RPG in AS400 for many years and I had never heard anything like this. Are you <b>absolutely sure</b> that the program you are running is the one corresponding to the RPG source? Do you have compiled it yourself? Perhaps you can check also the job stack while in the breakpoint and check the name and library (with SYSREQ, ...). Perhaps can exist another pgm with the same name performing a delete instead of a write... (I'm just guessing!). Check for pgm's with that name in all libraries (WRKOBJ *ALL/pgmname). Are you really sure that you do not have a trigger associate with that file? Are you journaling the file? If you are, you can display the "delete" journal entry and see what is the name of the program that is performing the write and the delete (take note: BOTH ENTRIES SHOULD BE IN THE JOURNAL). If do not have a journal, you can create one for the only purpose of finding what is going one with this issue. Do you have any OVRDBF? Do you use exit programs? I'll be listening for news in this. <b>Ok - here's the deal: I found another job running in a different subsystem that was picking up records and deleting them. It shouldn't have had that library in its list, but did. I'm still amazed at how quickly the update happened though because the other program was on a 5 second delay between I/O. I should have seen a record in the file before it was deleted, but never did.</b>