On 09/15/2014 11:12 AM, Tom Lane wrote:
>> Or are you proposing that JSONARRAY @> JSONARRAY should work differently
>> from ARRAY @> ARRAY?
>
> And no. It's a bug that jsonb array containment works differently from
> regular array containment. We understand the source of the bug, ie a
> mistaken optimization. I don't see why there's much need for discussion
> about anything except whether removing the optimization altogether
> (as Peter proposed) is the best fix, or whether we want to retain
> some weaker form of it.
Right, so I was just saying that after we fix this behavior, the
behavior of JSONARRAY @> JSONARRAY should be commented somewhere because
that comparison may not work the way users who are not long-time
postgres users expect. Heck, I've personally done very little ARRAY @>
ARRAY myself in 12 years of using PostgreSQL arrays; I had to test it to
verify the current behavior.
Not sure exactly where this note should go, mind you.
--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com