David Rowley <dgrowleyml@gmail.com> writes:
> On Thu, 30 Sept 2021 at 09:24, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> After further inspection, "what are we doing in jsonb_hash?" is
>> indeed a relevant question, but it seems like it's a type mismatch
>> not a nullness issue. EXPLAIN VERBOSE shows
> I think you're right here. It should be hashing text. That seems to
> be going wrong in check_memoizable() because it assumes it's always
> fine to use the left side's type of the OpExpr to figure out the hash
> function to use.
> Maybe we can cache the left and the right type's hash function and use
> the correct one in paraminfo_get_equal_hashops().
Um ... it seems to have correctly identified the cache key expressions,
so why isn't it just doing exprType on those? The jsonb_exists operator
seems entirely irrelevant here.
regards, tom lane