Re: future of serial and identity columns

On 10/4/22 09:41, Peter Eisentraut wrote:
> In PostgreSQL 10, we added identity columns, as an alternative to serial
> columns (since 6.something).  They mostly work the same.  Identity
> columns are SQL-conforming, have some more features (e.g., overriding
> clause), and are a bit more robust in schema management.  Some of that
> was described in [0].  AFAICT, there have been no complaints since that
> identity columns lack features or are somehow a regression over serial
> columns.
>
> But clearly, the syntax "serial" is more handy, and most casual examples
> use that syntax.  So it seems like we are stuck with maintaining these
> two variants in parallel forever.  I was thinking we could nudge this a
> little by remapping "serial" internally to create an identity column
> instead.  At least then over time, the use of the older serial
> mechanisms would go away.
>
> Note that pg_dump dumps a serial column in pieces (CREATE SEQUENCE +
> ALTER SEQUENCE ... OWNED BY + ALTER TABLE ... SET DEFAULT).  So if we
> did this, any existing databases would keep their old semantics, and
> those who really need it can manually create the old semantics as well.
>
> Attached is a demo patch how the implementation of this change would
> look like.  This creates a bunch of regression test failures, but
> AFAICT, those are mainly display differences and some very peculiar test
> setups that are intentionally examining some edge cases.  These would
> need to be investigated in more detail, of course.

I haven't tested the patch yet, just read it.

Is there any reason to use BY DEFAULT over ALWAYS?  I tend to prefer the
latter.