Re: pg_stats and range statistics

hi. I played around with the 2023-Apr 4 latest patch.

+        <literal>lower(ARRAY[numrange(1.1,2.2),numrange(3.3,4.4)])</literal>
should be
+        <literal>ranges_lower(ARRAY[numrange(1.1,2.2),numrange(3.3,4.4)])</literal>

+        <literal>upper(ARRAY[numrange(1.1,2.2),numrange(3.3,4.4)])</literal>
should be
+        <literal>ranges_upper(ARRAY[numrange(1.1,2.2),numrange(3.3,4.4)])</literal>

https://www.postgresql.org/docs/current/catalog-pg-type.html
there is no association between numrange and their base type numeric.
so for template: anyarray ranges_lower(anyarray). I don't think we can
input numrange array and return a numeric array.

https://www.postgresql.org/docs/current/extend-type-system.html#EXTEND-TYPES-POLYMORPHIC
>> When the return value of a function is declared as a polymorphic type, there must be at least one argument position that is also >> polymorphic, and the actual data type(s) supplied for the polymorphic arguments determine the actual result type for that call.


regression=# select
ranges_lower(ARRAY[numrange(1.1,2.2),numrange(3.3,4.4),
numrange(5.5,6.6)]);
 ranges_lower
---------------
 {1.1,3.3,5.5}
(1 row)
regression=# \gdesc
    Column    |    Type
--------------+------------
 ranges_lower | numrange[]
(1 row)

I don't think you can cast literal ' {1.1,3.3,5.5}' to numrange[].