Re: [HACKERS] Behavior of GENERATED columns per SQL2003 - Mailing list pgsql-patches

From Tom Lane
Subject Re: [HACKERS] Behavior of GENERATED columns per SQL2003
Date
Msg-id 1347.1178636561@sss.pgh.pa.us
Whole thread Raw
In response to Re: [HACKERS] Behavior of GENERATED columns per SQL2003  (Zoltan Boszormenyi <zb@cybertec.at>)
List pgsql-patches
Zoltan Boszormenyi <zb@cybertec.at> writes:
> Tom Lane �rta:
>> * There is no GENERATED BY DEFAULT AS (expr) in the spec.

> Yes, and because it already exists and called DEFAULT.

No, plain DEFAULT isn't a substitute, because it can't reference the
values of other columns.  I think there is a use-case for an overridable
GENERATED expression, but won't press the point right now.

>> Is anyone familiar with a database that implements SQL-spec generated
>> columns?  Do they actually store the columns?

> David Fuhry answered this question with the advertised intention.

Yeah, that's good to know --- the text of the spec is quite unhelpful
about telling you what they intended ...

            regards, tom lane

pgsql-patches by date:

Previous
From: Zoltan Boszormenyi
Date:
Subject: Re: [HACKERS] Behavior of GENERATED columns per SQL2003
Next
From: Ale Raza
Date:
Subject: Patch: Allocation of "kind" codes for spatial type.