Re: Making jsonb_agg() faster

From: Chao Li <li(dot)evan(dot)chao(at)gmail(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: jian he <jian(dot)universality(at)gmail(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: Making jsonb_agg() faster
Date: 2025-08-25 09:19:34
Message-ID: 68A1061C-81A5-46AA-AFA9-48294908254C@gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

> On Aug 23, 2025, at 03:11, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
>
> v2-0001 takes care of that, and also adopts your suggestion in [1]
> about not using two calls of pushJsonbValueScalar where one would do.
> I also did a bit more micro-optimization in appendKey, appendValue,
> appendElement to avoid redundant copying, because perf testing showed
> that appendElement is still a hot-spot for jsonb_agg. Patches 0002
> and 0003 are unchanged.
>

A small comment. In most of places, JsonbInState is initialized as “JsonbInState ps = {0};”, which is the preferred C style. There are some other places still using memset() to zero out JsonbInState variables. Maybe we can change all memset() usages to “{0}”.

I did a search and found out all memset to change:

diff --git a/src/backend/utils/adt/jsonb.c b/src/backend/utils/adt/jsonb.c
index 0c964052958..9ca501a7a6e 100644
--- a/src/backend/utils/adt/jsonb.c
+++ b/src/backend/utils/adt/jsonb.c
@@ -240,10 +240,10 @@ static inline Datum
jsonb_from_cstring(char *json, int len, bool unique_keys, Node *escontext)
{
JsonLexContext lex;
- JsonbInState state;
+ JsonbInState state = {0};
JsonSemAction sem;

- memset(&state, 0, sizeof(state));
+ //memset(&state, 0, sizeof(state));
memset(&sem, 0, sizeof(sem));
makeJsonLexContextCstringLen(&lex, json, len, GetDatabaseEncoding(), true);

@@ -1147,9 +1147,9 @@ to_jsonb(PG_FUNCTION_ARGS)
Datum
datum_to_jsonb(Datum val, JsonTypeCategory tcategory, Oid outfuncoid)
{
- JsonbInState result;
+ JsonbInState result = {0};

- memset(&result, 0, sizeof(JsonbInState));
+ //memset(&result, 0, sizeof(JsonbInState));

datum_to_jsonb_internal(val, false, &result, tcategory, outfuncoid,
false);
@@ -1162,7 +1162,7 @@ jsonb_build_object_worker(int nargs, const Datum *args, const bool *nulls, const
bool absent_on_null, bool unique_keys)
{
int i;
- JsonbInState result;
+ JsonbInState result = {0};

if (nargs % 2 != 0)
ereport(ERROR,
@@ -1172,7 +1172,7 @@ jsonb_build_object_worker(int nargs, const Datum *args, const bool *nulls, const
errhint("The arguments of %s must consist of alternating keys and values.",
"jsonb_build_object()")));

- memset(&result, 0, sizeof(JsonbInState));
+ //memset(&result, 0, sizeof(JsonbInState));

pushJsonbValue(&result, WJB_BEGIN_OBJECT, NULL);
result.parseState->unique_keys = unique_keys;
@@ -1232,9 +1232,9 @@ jsonb_build_object(PG_FUNCTION_ARGS)
Datum
jsonb_build_object_noargs(PG_FUNCTION_ARGS)
{
- JsonbInState result;
+ JsonbInState result = {0};

- memset(&result, 0, sizeof(JsonbInState));
+ //memset(&result, 0, sizeof(JsonbInState));

pushJsonbValue(&result, WJB_BEGIN_OBJECT, NULL);
pushJsonbValue(&result, WJB_END_OBJECT, NULL);
@@ -1293,9 +1293,9 @@ jsonb_build_array(PG_FUNCTION_ARGS)
Datum
jsonb_build_array_noargs(PG_FUNCTION_ARGS)
{
- JsonbInState result;
+ JsonbInState result = {0};

- memset(&result, 0, sizeof(JsonbInState));
+ //memset(&result, 0, sizeof(JsonbInState));

pushJsonbValue(&result, WJB_BEGIN_ARRAY, NULL);
pushJsonbValue(&result, WJB_END_ARRAY, NULL);
@@ -1321,9 +1321,9 @@ jsonb_object(PG_FUNCTION_ARGS)
int in_count,
count,
i;
- JsonbInState result;
+ JsonbInState result = {0};

- memset(&result, 0, sizeof(JsonbInState));
+ //memset(&result, 0, sizeof(JsonbInState));

pushJsonbValue(&result, WJB_BEGIN_OBJECT, NULL);

@@ -1425,9 +1425,9 @@ jsonb_object_two_arg(PG_FUNCTION_ARGS)
int key_count,
val_count,
i;
- JsonbInState result;
+ JsonbInState result = {0};

- memset(&result, 0, sizeof(JsonbInState));
+ //memset(&result, 0, sizeof(JsonbInState));

pushJsonbValue(&result, WJB_BEGIN_OBJECT, NULL);

diff --git a/src/backend/utils/adt/jsonpath_exec.c b/src/backend/utils/adt/jsonpath_exec.c
index f0d2378a8cc..4305623b979 100644
--- a/src/backend/utils/adt/jsonpath_exec.c
+++ b/src/backend/utils/adt/jsonpath_exec.c
@@ -2872,7 +2872,7 @@ executeKeyValueMethod(JsonPathExecContext *cxt, JsonPathItem *jsp,
{
JsonBaseObjectInfo baseObject;
JsonbValue obj;
- JsonbInState ps;
+ JsonbInState ps = {0};
Jsonb *jsonb;

if (tok != WJB_KEY)
@@ -2886,7 +2886,7 @@ executeKeyValueMethod(JsonPathExecContext *cxt, JsonPathItem *jsp,
tok = JsonbIteratorNext(&it, &val, true);
Assert(tok == WJB_VALUE);

- memset(&ps, 0, sizeof(ps));
+ //memset(&ps, 0, sizeof(ps));

pushJsonbValue(&ps, WJB_BEGIN_OBJECT, NULL);

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2025-08-25 09:36:28 Re: Generate GUC tables from .dat file
Previous Message Ashutosh Bapat 2025-08-25 09:10:42 Calling PGReserveSemaphores() from CreateOrAttachShmemStructs