type cache for concat functions
Hi
Now concat is 2x times slower than || operator. With cached FmgrInfo for
output function it will be only 5%.
Regards
Pavel
Attachments:
faster-concat.patchtext/x-patch; charset=US-ASCII; name=faster-concat.patchDownload
diff --git a/src/backend/utils/adt/varlena.c b/src/backend/utils/adt/varlena.c
index be399f4..d3f04f8 100644
--- a/src/backend/utils/adt/varlena.c
+++ b/src/backend/utils/adt/varlena.c
@@ -4718,6 +4718,28 @@ string_agg_finalfn(PG_FUNCTION_ARGS)
PG_RETURN_NULL();
}
+static void
+rebuildConcatCache(FmgrInfo *typcache, FunctionCallInfo fcinfo, int argidx)
+{
+ int i;
+
+ Assert(typcache != NULL);
+
+ for (i = argidx; i < PG_NARGS(); i++)
+ {
+ Oid valtype;
+ Oid typOutput;
+ bool typIsVarlena;
+
+ valtype = get_fn_expr_argtype(fcinfo->flinfo, i);
+ if (!OidIsValid(valtype))
+ elog(ERROR, "could not determine data type of concat() input");
+
+ getTypeOutputInfo(valtype, &typOutput, &typIsVarlena);
+ fmgr_info_cxt(typOutput, &typcache[i-argidx], fcinfo->flinfo->fn_mcxt);
+ }
+}
+
/*
* Implementation of both concat() and concat_ws().
*
@@ -4733,6 +4755,7 @@ concat_internal(const char *sepstr, int argidx,
StringInfoData str;
bool first_arg = true;
int i;
+ FmgrInfo *typcache;
/*
* concat(VARIADIC some-array) is essentially equivalent to
@@ -4772,14 +4795,20 @@ concat_internal(const char *sepstr, int argidx,
/* Normal case without explicit VARIADIC marker */
initStringInfo(&str);
+ typcache = (FmgrInfo *) fcinfo->flinfo->fn_extra;
+ if (typcache == NULL)
+ {
+ fcinfo->flinfo->fn_extra = MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ PG_NARGS() * sizeof(FmgrInfo));
+ typcache = (FmgrInfo *) fcinfo->flinfo->fn_extra;
+ rebuildConcatCache(typcache, fcinfo, argidx);
+ }
+
for (i = argidx; i < PG_NARGS(); i++)
{
if (!PG_ARGISNULL(i))
{
Datum value = PG_GETARG_DATUM(i);
- Oid valtype;
- Oid typOutput;
- bool typIsVarlena;
/* add separator if appropriate */
if (first_arg)
@@ -4787,13 +4816,8 @@ concat_internal(const char *sepstr, int argidx,
else
appendStringInfoString(&str, sepstr);
- /* call the appropriate type output function, append the result */
- valtype = get_fn_expr_argtype(fcinfo->flinfo, i);
- if (!OidIsValid(valtype))
- elog(ERROR, "could not determine data type of concat() input");
- getTypeOutputInfo(valtype, &typOutput, &typIsVarlena);
appendStringInfoString(&str,
- OidOutputFunctionCall(typOutput, value));
+ OutputFunctionCall(&typcache[i-argidx], value));
}
}
On 20 May 2017 at 10:03, Pavel Stehule <pavel.stehule@gmail.com> wrote:
Now concat is 2x times slower than || operator. With cached FmgrInfo for
output function it will be only 5%.
Looks nice and does what's expected (what else one may need from a Saturday
evening). I just can mention that from what I see the previous version of
`concat` is slower the more arguments are getting involved, so looks like it
can be more than 2x.
Also, it was a little bit confusing to see that the function `concat`
behaves
differently from operator `||` in terms of performance. When you're looking
at
the code it's becoming obvious, but I couldn't find any mention about that
in
the documentation.
2017-06-18 0:50 GMT+02:00 Dmitry Dolgov <9erthalion6@gmail.com>:
On 20 May 2017 at 10:03, Pavel Stehule <pavel.stehule@gmail.com> wrote:
Now concat is 2x times slower than || operator. With cached FmgrInfo for
output function it will be only 5%.Looks nice and does what's expected (what else one may need from a Saturday
evening). I just can mention that from what I see the previous version of
`concat` is slower the more arguments are getting involved, so looks like
it
can be more than 2x.Also, it was a little bit confusing to see that the function `concat`
behaves
differently from operator `||` in terms of performance. When you're
looking at
the code it's becoming obvious, but I couldn't find any mention about that
in
the documentation.
There was few steps how to evaluate expressions more faster, but no step
how to evaluate faster polymorphic "any" functions. A detecting of used
argument type and looking to type cache are expensive. These functions was
designed for string processing in error messages, and the performance was
not too important. Now the concat function can be used often as replacement
of || operator when you do migration from Oracle due similar behave. And
the performance is more interesting now.
I think so can be interesting to allow some activity in analyzing/planning
time and prexec stage to user function developers. Some is possible via
planning time hooks, but for controlling few functions it is not user
friendly - and there are no space for data from this stages.
Regards
Pavel
Hi Pavel,
I tried applying your patch, it applies and compiles fine, check and
checkworld pass.
I ran a simple performance test, select
concat(generate_series(1,100000), ... [x5 total]) vs select
generate_series(1,100000)::text || ... .
Operator || runs in 60 ms, while unpatched concat takes 90 and patched
-- 55 ms.
About the code:
* There seems to be an extra tab here:
FmgrInfo *typcache;
It's not aligned with the previous declaration with tab size 4.
* You could allocate the cache and store it into the context inside
rebuildConcatCache. It would return return the cache pointer, and the
calling code would look cleaner -- just one line. This is a matter of
taste though.
* The nearby functions use snake_case, so it should be
rebuild_concat_cache too.
Overall the patch looks good to me.
--
Alexander Kuzmenkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Hi
2017-08-24 19:24 GMT+02:00 Alexander Kuzmenkov <a.kuzmenkov@postgrespro.ru>:
Hi Pavel,
I tried applying your patch, it applies and compiles fine, check and
checkworld pass.I ran a simple performance test, select concat(generate_series(1,100000),
... [x5 total]) vs select generate_series(1,100000)::text || ... .
Operator || runs in 60 ms, while unpatched concat takes 90 and patched --
55 ms.About the code:
* There seems to be an extra tab here:
FmgrInfo *typcache;
It's not aligned with the previous declaration with tab size 4.* You could allocate the cache and store it into the context inside
rebuildConcatCache. It would return return the cache pointer, and the
calling code would look cleaner -- just one line. This is a matter of taste
though.* The nearby functions use snake_case, so it should be
rebuild_concat_cache too.
all should be fixed.
Regards
Pavel
Show quoted text
Overall the patch looks good to me.
--
Alexander Kuzmenkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company
Attachments:
faster-concat-2.patchtext/x-patch; charset=US-ASCII; name=faster-concat-2.patchDownload
diff --git a/src/backend/utils/adt/varlena.c b/src/backend/utils/adt/varlena.c
index ebfb823fb8..24d7512320 100644
--- a/src/backend/utils/adt/varlena.c
+++ b/src/backend/utils/adt/varlena.c
@@ -4734,6 +4734,38 @@ string_agg_finalfn(PG_FUNCTION_ARGS)
}
/*
+ * Prepare cache with typeOutput fmgr info for any argument of
+ * concat like function.
+ */
+static FmgrInfo *
+build_concat_typcache(FunctionCallInfo fcinfo, int argidx)
+{
+ FmgrInfo *typcache;
+ int i;
+
+ typcache = MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
+ PG_NARGS() * sizeof(FmgrInfo));
+
+ for (i = argidx; i < PG_NARGS(); i++)
+ {
+ Oid valtype;
+ Oid typOutput;
+ bool typIsVarlena;
+
+ valtype = get_fn_expr_argtype(fcinfo->flinfo, i);
+ if (!OidIsValid(valtype))
+ elog(ERROR, "could not determine data type of concat() input");
+
+ getTypeOutputInfo(valtype, &typOutput, &typIsVarlena);
+ fmgr_info_cxt(typOutput, &typcache[i-argidx], fcinfo->flinfo->fn_mcxt);
+ }
+
+ fcinfo->flinfo->fn_extra = typcache;
+
+ return typcache;
+}
+
+/*
* Implementation of both concat() and concat_ws().
*
* sepstr is the separator string to place between values.
@@ -4748,6 +4780,7 @@ concat_internal(const char *sepstr, int argidx,
StringInfoData str;
bool first_arg = true;
int i;
+ FmgrInfo *typcache;
/*
* concat(VARIADIC some-array) is essentially equivalent to
@@ -4787,14 +4820,15 @@ concat_internal(const char *sepstr, int argidx,
/* Normal case without explicit VARIADIC marker */
initStringInfo(&str);
+ typcache = (FmgrInfo *) fcinfo->flinfo->fn_extra;
+ if (typcache == NULL)
+ typcache = build_concat_typcache(fcinfo, argidx);
+
for (i = argidx; i < PG_NARGS(); i++)
{
if (!PG_ARGISNULL(i))
{
Datum value = PG_GETARG_DATUM(i);
- Oid valtype;
- Oid typOutput;
- bool typIsVarlena;
/* add separator if appropriate */
if (first_arg)
@@ -4802,13 +4836,8 @@ concat_internal(const char *sepstr, int argidx,
else
appendStringInfoString(&str, sepstr);
- /* call the appropriate type output function, append the result */
- valtype = get_fn_expr_argtype(fcinfo->flinfo, i);
- if (!OidIsValid(valtype))
- elog(ERROR, "could not determine data type of concat() input");
- getTypeOutputInfo(valtype, &typOutput, &typIsVarlena);
appendStringInfoString(&str,
- OidOutputFunctionCall(typOutput, value));
+ OutputFunctionCall(&typcache[i-argidx], value));
}
}
The following review has been posted through the commitfest application:
make installcheck-world: tested, passed
Implements feature: tested, passed
Spec compliant: tested, passed
Documentation: tested, passed
Looks good to me.
The new status of this patch is: Ready for Committer
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
2017-09-19 12:18 GMT+02:00 Alexander Kuzmenkov <a.kuzmenkov@postgrespro.ru>:
The following review has been posted through the commitfest application:
make installcheck-world: tested, passed
Implements feature: tested, passed
Spec compliant: tested, passed
Documentation: tested, passedLooks good to me.
The new status of this patch is: Ready for Committer
Thank you very much
Pavel
Show quoted text
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Pavel Stehule <pavel.stehule@gmail.com> writes:
[ faster-concat-2.patch ]
Pushed with some cosmetic adjustments (mostly better comments).
regards, tom lane
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers