Re^2: RFC: Placeholder creation for SQL statements

Native Postgresql uses the dollar-integer syntax for placeholders in prepared queries and doesn't support question marks at all, that is emulated by DBD::Pg. That is why you have to take care when you use question marks in SQL comments or as an operator, you have to escape them.

Comment on Re^2: RFC: Placeholder creation for SQL statements

Replies are listed 'Best First'.
Re^3: RFC: Placeholder creation for SQL statements by erix (Prior) on Aug 29, 2022 at 13:36 UTC
for ?-containing operators, of course. But I can't find a way via DBI to let a postgres server complain about a questionmark in comments. (tried server versions 14, 15, 16). This kinda makes sense because '--'-style comments used to be removed, although '/* */'-style comments were kept. (btw, this will be changed (to both included) in next Pg version, 15) (It would be another good reason to use `$dbh->{pg_placeholder_dollaronly} = 1;`. To me the operator-with-questionmark case is reason enough to use it)	[reply] [d/l]
Re^4: RFC: Placeholder creation for SQL statements by kikuchiyo (Hermit) on Aug 29, 2022 at 14:34 UTC
Hmm, indeed, it seems that DBD::Pg handles question marks in comments correctly now (3.16.0). <Reveal this spoiler or all in this thread> I distinctly remember that this used to be a problem a few years ago. But you're right about the operator-with-questionmark case: there is actually a "?" operator in postgres (to test for a key in a jsonb object), and you have to escape it if you want to use it in a query with DBD::Pg (which makes the SQL incompatible if you want to reuse it with e.g. raw psql or Go/sqlc).	[reply] [d/l]