What order are you issuing EXFMT CHAIN or READ UPDATE ? Phil ---------------------- yes.. exactly… .. what? Performs an update .. with a native I/O UPDATE command? F-spec doesn't have a prefix keyword You are saying that in debug you do an EVAL var1 just before the udpate and it's good What happens if you do EVAL var1 just after the update? Phil ---------- Are you updating the physical file or through a logical file? Phil ------------------------------ with native I/O UPDATE command? No prefix on the F-Spec No rename of fields It's been compiled over the PF with the field order of events chain/read exfmt dspfrecord update record Phil ---------------- "What happens if you do EVAL var1 just after the update? - it looks good too." Then it cannot be recognizing the var1 as a part of the record that is updated. Phil ------------------------- F-spec 'E'xternally defined?? or 'F'ixed format? Phil ------------------------------ If the sequence is to do an EXFMT with that variable name defined in the display format, <b>then</b> CHAIN / READ a record from the file with the same variable name defined in the record format, the contents of the variable are being replaced by the last operation. Either load the variable from the file prior to displaying it on the screen (if the current value needs displaying) or use a different name for the variable on the display format. Splat ------------------------------ Phil, My apologies. I didn't see your comment about the sequence of events until after I'd posted the above. Splat ------------- Don't apologize for trying to help..the key was his response -- the value was in the field even after the update .. I don't think his program sees the field in the file .. thus the question about Fixed format ... Phil ------------------ Please tell the problem/solution .. it might help others. Phil

What order are you issuing EXFMT CHAIN or READ UPDATE ? Phil ---------------------- yes.. exactly… .. what? Performs an update .. with a native I/O UPDATE command? F-spec doesn't have a prefix keyword You are saying that in debug you do an EVAL var1 just before the udpate and it's good What happens if you do EVAL var1 just after the update? Phil ---------- Are you updating the physical file or through a logical file? Phil ------------------------------ with native I/O UPDATE command? No prefix on the F-Spec No rename of fields It's been compiled over the PF with the field order of events chain/read exfmt dspfrecord update record Phil ---------------- "What happens if you do EVAL var1 just after the update? - it looks good too." Then it cannot be recognizing the var1 as a part of the record that is updated. Phil ------------------------- F-spec 'E'xternally defined?? or 'F'ixed format? Phil ------------------------------ If the sequence is to do an EXFMT with that variable name defined in the display format, <b>then</b> CHAIN / READ a record from the file with the same variable name defined in the record format, the contents of the variable are being replaced by the last operation. Either load the variable from the file prior to displaying it on the screen (if the current value needs displaying) or use a different name for the variable on the display format. Splat ------------------------------ Phil, My apologies. I didn't see your comment about the sequence of events until after I'd posted the above. Splat ------------- Don't apologize for trying to help..the key was his response -- the value was in the field even after the update .. I don't think his program sees the field in the file .. thus the question about Fixed format ... Phil ------------------ Please tell the problem/solution .. it might help others. Phil