On Thu, Feb 9, 2023 at 9:28 AM Alban Hertroys <haramrae@xxxxxxxxx> wrote:
> On 9 Feb 2023, at 16:41, Dominique Devienne <ddevienne@xxxxxxxxx> wrote:
>
> Hi. We are implementing an API which takes a list of row keys, and must return info about those rows. To implement that efficiently, in as few round-trips as possible, we bind a (binary) array of keys (ints, uuids, or strings) and that works great, but only if the key is a scalar one.
>
> Now we'd like to do the same for composite keys, and I don't know how to do that.
> Is it possible? Could someone please help out or demo such a thing?
> We are doing it in C++ using libpq, but a pure SQL or PL/pgSQL demo would still help (I think).
This works:
=> select (1, 'one'::text) in ((1, 'two'::text), (2, 'one'::text), (1, 'one'::text), (2, 'two'::text));
?column?
----------
t
(1 row)
But you cannot write the right-side of the IN as a single parameter which seems to be the primary constraint trying to be conformed to.
David J.
