Re: IDENTITY/GENERATED v36 Re: Final version of IDENTITY/GENERATED patch - Mailing list pgsql-patches

From Tom Lane
Subject Re: IDENTITY/GENERATED v36 Re: Final version of IDENTITY/GENERATED patch
Date
Msg-id 18616.1175747962@sss.pgh.pa.us
Whole thread Raw
In response to Re: IDENTITY/GENERATED v36 Re: Final version of IDENTITY/GENERATED patch  (Zoltan Boszormenyi <zb@cybertec.at>)
List pgsql-patches
Zoltan Boszormenyi <zb@cybertec.at> writes:
> Tom Lane �rta:
>> Zoltan Boszormenyi <zb@cybertec.at> writes:
>>> Yes. Plain SERIALs can be updated with given values
>>> whereas IDENTITY columns cannot.
>>
>> Really?  How is pg_dump going to deal with that?

> It emits ALTER TABLE ... SET GENERATED AS IDENTITY
> after ALTER SEQUENCE OWNED BY and if any of the
> columns is IDENTITY or GENERATED then it emits
> COPY OVERRIDING SYSTEM VALUE in my patch already.

And you fail to see the irony in that?  You might as well just
admit that it's OK to update an identity column.

            regards, tom lane

pgsql-patches by date:

Previous
From: Zoltan Boszormenyi
Date:
Subject: Re: IDENTITY/GENERATED v36 Re: Final version of IDENTITY/GENERATED patch
Next
From: "Simon Riggs"
Date:
Subject: Re: [HACKERS] Full page writes improvement, code update