Horton

Re: Making jsonb_agg() faster

Tom Lane <tgl@sss.pgh.pa.us>

View thread
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Chao Li <li.evan.chao@gmail.com>
Cc: jian he <jian.universality@gmail.com>, pgsql-hackers@lists.postgresql.org
Date: 2025-12-06T19:00:48Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix jsonb_object_agg crash after eliminating null-valued pairs.

  2. Micro-optimize datatype conversions in datum_to_jsonb_internal.

  3. Remove fundamentally-redundant processing in jsonb_agg() et al.

  4. Revise APIs for pushJsonbValue() and associated routines.

  5. Add a macro for the declared typlen of type timetz.

Chao Li <li.evan.chao@gmail.com> writes:
> On Dec 6, 2025, at 07:14, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> I'd kind of like to get this pushed soon, because it keeps getting
>> sideswiped ... does anyone have further comments?

> My only nit commit is still about the hard-coded 12:
> I commented this before and you explained. But I still think it may deserve a comment for why 12 is here, otherwise future reader may also get the same confusion as when I first time read this code.

I've been thinking of that as an independent issue.  But what I'm
inclined to do is add a symbol to date.h, say like

diff --git a/src/include/utils/date.h b/src/include/utils/date.h
index 7316ac0ff17..2aca785b65d 100644
--- a/src/include/utils/date.h
+++ b/src/include/utils/date.h
@@ -30,6 +30,14 @@ typedef struct
 	int32		zone;			/* numeric time zone, in seconds */
 } TimeTzADT;
 
+/*
+ * sizeof(TimeTzADT) will be 16 on most platforms due to alignment padding.
+ * However, timetz's typlen is 12 according to pg_type.  In most places
+ * we can get away with using sizeof(TimeTzADT), but where it's important
+ * to match the declared typlen, use TIMETZ_TYPLEN.
+ */
+#define TIMETZ_TYPLEN		12
+
 /*
  * Infinity and minus infinity must be the max and min values of DateADT.
  */

and then use that.  (I poked around in other code using TimeTzADT,
and could not find any other places where we have hard-wired "12",
which seems a bit surprising perhaps.  But pretty much all the
references to sizeof(TimeTzADT) are in palloc's, where it's fine.)

			regards, tom lane