| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | "David G(dot) Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com> |
| Cc: | exe-dealer <exe-dealer(at)yandex(dot)ru>, "pgsql-bugs(at)lists(dot)postgresql(dot)org" <pgsql-bugs(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: jsonb subscript operator returns null when key is fetched from table |
| Date: | 2022-12-12 19:09:21 |
| Message-ID: | 2035559.1670872161@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
"David G. Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com> writes:
> Someone more experienced than I will need to actually do the debugging work
> though; hoping this reply will be seen by one of them.
Meh. Somebody was cavalier about detoasting the subscript:
jbvp = getKeyJsonValueFromContainer(container,
VARDATA(path[i]),
VARSIZE(path[i]) - VARHDRSZ,
NULL);
Neither VARDATA nor VARSIZE are valid to apply to an arbitrary datum.
Apparently, nobody has ever tested this with a subscript value fetched
straight from disk :-(. The given example accidentally works with
a NOT MATERIALIZED CTE because the whole thing gets stepwise folded
to constants. Likewise, the expression with || accidentally works
because concatenation always yields a non-toasted result.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Peter Eisentraut | 2022-12-12 19:16:32 | Re: BUG #17434: CREATE/DROP DATABASE can be executed in the same transaction with other commands |
| Previous Message | David G. Johnston | 2022-12-12 18:51:19 | Re: jsonb subscript operator returns null when key is fetched from table |