To prevent record locks or lock-related errors when using CHAIN, UPDATE, or DELETE in RPGLE, you can use a combination of safe access methods, error handling, and access strategies. Here’s how to do it effectively.
✅ 1. Use N Operation Extender (CHAIN(N)): No Lock Read
Use this if you only want to read the record without locking it.
Chain (N: CustomerID) Customers;
If %Found(Customers);
Dsply 'Read without locking.';
EndIf;
🔸 Best for display-only or logging purposes.
✅ 2. Use INFSR or MONITOR for Error Handling
Use MONITOR / ON-ERROR blocks or %ERROR with %STATUS to catch locks safely.
Example:
Monitor;
Chain (CustomerID) Customers;
If %Found(Customers);
Name = 'Safe Update';
Update Customers;
EndIf;
On-Error 1218;
Dsply 'Record is locked, skipping update.';
EndMon;
✅ 3. Avoiding Locking Conflicts
A. Use CHAIN(N) followed by UPDATE only if needed:
Chain (N: CustomerID) Customers; // No lock here
If %Found(Customers);
// Do something, then lock only when updating
Chain (CustomerID) Customers; // With lock
If Not %Error;
Name = 'Final Update';
Update Customers;
ElseIf %Status = 1218;
Dsply 'Locked during update.';
EndIf;
EndIf;
This limits the time the record is locked.
✅ 4. Use Record-Level Lock Wait Option in F-spec
In your DCL-F file declaration:
Dcl-F Customers Usage(*Update) Keyed LockWait(0);
| LockWait(x) | Behavior |
|---|---|
LockWait(0) |
Do not wait if record is locked |
LockWait(5) |
Wait 5 seconds before giving up |
LockWait(*NoWait) |
Synonym for 0-second wait |
LockWait(*MaxWait) |
Wait indefinitely |
If lock fails, use %ERROR and %STATUS to handle gracefully.
✅ 5. Avoid READ or CHAIN inside tight loops on busy files
If the file is heavily used, reduce lock time and avoid nested CHAIN/UPDATE inside long loops. Instead, load data in arrays or use embedded SQL cursors with WITH RS USE AND KEEP SHARE LOCKS.
✅ 6. Use *IN indicators for legacy-style lock checks
For older-style RPG:
C Chain(E) KeyField MyFile 90
C If *IN90
C // Locked or not found
C EndIf