Zoltan Boszormenyi írta:
>> The cost I am thinking now is an extra heap_update()
>> after heap_insert() without generating the identity value
>> and inserting index tuples to indexes that doesn't
>> contain the identity column.
>
> And as far as I tested the current state, there is no cost
> if you don't use GENERATED or IDENTITY.
> The extra heap_update() would be performed only
> if you have an IDENTITY colum.
The modification I imagined is actually working:
- skip indexes using the identity columns
- do a simple_heap_update() after all other columns are assigned and index tuples are inserted
- do ExecInsertIndexTuples() on indexes referencing the IDENTITY column
However, I get warning messages like:
WARNING: detected write past chunk end in ExecutorState 0xaaff68
How can I prevent them?
Best regards,
Zoltán Böszörményi