Re: Proposed patch: make SQL interval-literal syntax work per spec

Tom Lane wrote:
> Ron Mayer <rm_pg(at)cheapcomplexdevices(dot)com> writes:
>> Back a while ago (2003) there was some talk about replacing
>> some of the non-standard extensions with shorthand forms of
>> intervals with ISO 8601 intervals that have a similar but
>> not-the-same shorthand.
>
> I think *replacement* would be a hard sell, as that would tick off all
> the existing users ;-). Now it seems like being able to accept either

I originally submitted a patch that supported both, and I think
you suggested replacing on the grounds that the old one was
never documented,

http://archives.postgresql.org/pgsql-patches/2003-09/msg00134.php
"If we're going to support the real ISO spec, I'd suggest ripping
out any not-quite-there variant.

http://archives.postgresql.org/pgsql-patches/2003-09/msg00121.php
"I doubt anyone is using it, because it's completely undocumented."

On the other hand, the company I was at was indeed originally
using it, so I prefer that it stay in as well. Perhaps if
there's a way to mark them as deprecated and post warnings in
the log file if they're used. I think they should be
removed eventually in a few releases, because they're quite
confusing as they stand:

Interval ISO Postgres
8601 shorthand
-----------------------------------------------------
'1 year 1 minute' 'P1YT1M' '1Y1M'
'1 year 1 month' 'P1Y1M' N/A

> the 8601 syntax or the existing syntaxes on input wouldn't be tough
> at all, if you insist on the P prefix to distinguish; so that end of

ISO 8601 seems to me to require the P, so I think we would.

> it should be easy enough. On the output side, seems like a GUC variable
> is the standard precedent here. I'd still vote against overloading
> DateStyle --- it does too much already --- but a separate variable for
> interval style wouldn't bother me. In fact, given that we are now
> somewhat SQL-compliant on interval input, a GUC that selected
> PG traditional, SQL-standard, or ISO 8601 interval output format seems
> like it could be a good idea.

Great. I'm bringing my patch up-to-date with CVS now
and adding the separate GUC.