On Tue, May 8, 2018 at 7:13 PM, Peter Eisentraut
<peter.eisentraut@2ndquadrant.com> wrote:
> On 5/8/18 11:31, Tom Lane wrote:
>> Paul Howells <paul.steven.howells@gmail.com> writes:
>>> Has there been or is there any current effort to implement SQL:2011
>>> valid-time support in Postgres?
>>
>> Searching the archives, I can only find "valid-time" appearing in these
>> threads related to temporal query processing:
>>
>> https://www.postgresql.org/message-id/flat/CALNdv1h7TUP24Nro53KecvWB2kwA67p%2BPByDuP6_1GeESTFgSA%40mail.gmail.com
>
> I think those are operators that work on top of having the valid-time
> available, but don't do anything about making that time available.
I think it is the reverse actually: these operators are more
general-purpose than valid time and could be used internally to
implement it, much like Postgres uses constraint triggers to implement
foreign keys. You could even leave them unexposed to the user (as
suggested in their papers), although personally I would love to have
access to them. Anyway, I think supporting valid-time tables (one way
or the other) in Postgres would be a tremendous addition.
I'm traveling and can't write much more at the moment, but I'll try to
reply more fully in a week or two. I'm happy to offer to help with
this. My only Postgres contribution so far was a tiny thing about
uuids and btree_gist, but I've spent a lot of time reading the source
code to build various extensions, and temporal databases has been an
interest of mine for several years.
Yours,
Paul