-
Task
-
Resolution: Done
-
Major
-
1.7.0.CR1
-
None
-
False
-
False
-
Undefined
-
E.g. here for MySQL, but it's the same for other connectors. The docs say that the regular expression is used to identify the key columns, but actually, the regular expression is used to match the table(s), whereas the corresponding key column(s) are given with their name.
See here for the actual semantics. Here's some valid examples:
# use column "pk1" as message key for table inventory.customer inventory.customers:pk1 # use columns "pk1 and pk2" as message key for table inventory.customer inventory.customers:pk1,pk2 # use columns "pk1 and pk2" as message key for table inventory.customer, "pk3" and "pk4" for tables "purchaseorders" in any schema inventory.customers:pk1,pk2;(.*).purchaseorders:pk3,pk4
