We have a number of optimizations of our work tables (pc_work equivalent) which require the use of oracle triggers to populate a few columns. The columns are non-nullable, and are not expected to be set/populated by Pega workflows. They are populated with Before Insert / Update trigger.
On issuing a Obj-Save / commit, we run into an issue where we get an error such as:
** trying to store an empty property MY_PROPERTY_NAME to a database column that does not accept nulls: code: <none> SQLState: <none> Message: <none>
This is not an Oracle error, but a Pega generated error because there is not an ORA error, nor is the query registered on database. There are also no relevant errors in the Pega logs except for EXCP0001 in the Pega Alerts indicating the issue above.
Question is: Without creating column-named properties, and defaulting their values (via Pega trigger, or something like NewDefaults), is there an ability to have Pega ignore column definition? Or at least allow inserts to non-nullable columns? It might be better to have this handled by database where we have a proper trackable error.
As of now, PRPC platform do not support the combination of NOT NULL and generated in the schema. PRPC wants to avoid “generating” queries and persisting user data that will appear to mismatch the requirements of your schema.