Re: Extended Statistics set/restore/clear functions.

On Mon, Nov 17, 2025 at 2:56 PM Michael Paquier <michael@paquier.xyz> wrote:
>
> On Fri, Nov 14, 2025 at 03:25:27PM +0900, Michael Paquier wrote:
> > Thanks for the new versions, I'll also look at all these across the
> > next couple of days.  Probably not at 0005~ for now.
>
> 0001 and 0002 from series v13 have been applied to change the output
> functions.
>
> And I have looked at 0003 in details for now.  Attached is a revised
> version for it, with many adjustments.  Some notes:
> - Many portions of the coverage were missing.  I have measured the
> coverage at 91% with the updated version attached.  This includes
> coverage for some error reporting, something that we rely a lot on for
> this code.
> - The error reports are made simpler, with the token values getting
> hidden.  While testing with some fancy values, I have actually noticed
> that the error handlings for the parsing of the int16 and int32 values
> were incorrect, the error reports used what the safe functions
> generated, not the reports from the data type.
> - Passing down arbitrary bytes sequences was leading to these bytes
> reported in the error outputs because we cared about the token values.
> I have added a few tests based on that for the code paths involved.
>
> There is an extra thing that bugs me as incorrect for the pg_ndistinct
> input, something I have not tackled myself yet.  Your patch checks
> that subsets of attributes are included in the longest set found, but
> it does not match the guarantees we have in mvndistinct.c: we have to
> check that *all* the combinations generated by generator_init() are
> satisfied based on the longest of attributes detected.  For example,
> this is thought as correct in the input function:
> SELECT '[{"attributes" : [-1,2], "ndistinct" : 1},
>          {"attributes" : [-1,2,3], "ndistinct" : 3}]'::pg_ndistinct;
>
> However it is obviously not correct as we are missing an element for
> the attributes [-1, 3].  The simplest solution would be to export the
> routines that generate the groups now in mvndistinct.c.  Also we
> should make sure that the number of elements in the arrays match with
> the number of groups we expect, not only the elements.  I don't think
> that we need to care much about the values, but we ought to provide
> stronger guarantees for the attributes listed in these elements.
>
> Except for this argument, the input of pg_ndistinct feels OK in terms
> of the guarantees that we'd want to enforce on an import.  The same
> argument applies in terms of attribute number guarantees for
> pg_dependencies, based on DependencyGenerator_init() & friends in
> dependencies.c.  Could you look at that?
>

hi.

NDistinctSemanticState, last element
NDIST_EXPECT_COMPLETE should follow with a comma, like:
``NDIST_EXPECT_COMPLETE, ``


Lots of tests use pg_input_error_info, which is good.
since currently only pg_input_error_info, pg_input_is_valid
NDistinctParseState.escontext is not NULL.


+ /*
+ * If escontext already set, just use that. Anything else is a generic
+ * JSON parse error.
+ */
+ if (!SOFT_ERROR_OCCURRED(parse_state.escontext))
+ errsave(parse_state.escontext,
+ errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
+ errmsg("malformed pg_ndistinct: \"%s\"", str),
+ errdetail("Must be valid JSON."));
seems no coverage for this.

segfault:
SELECT '[{"attributes" : [1,2,3,4,5,67,6,7,8], "ndistinct" : 4}]'::pg_ndistinct;

because src/backend/statistics/mvdistinct.c line: 310, Assert
Assert((item->nattributes >= 2) && (item->nattributes <= STATS_MAX_DIMENSIONS));


+SELECT '[{"attributes" : [2], "ndistinct" : 4}]'::pg_ndistinct;
+ERROR:  malformed pg_ndistinct: "[{"attributes" : [2], "ndistinct" : 4}]"
+LINE 1: SELECT '[{"attributes" : [2], "ndistinct" : 4}]'::pg_ndistin...
+               ^
+DETAIL:  The "ndistinct" key must contain an array of at least two attributes.

here, it should be
+DETAIL:  The "attributes" key must contain an array of at least two attributes.
?

--
jian
https://www.enterprisedb.com/