I use a statistical tracking software that allows user to CREATE their own custom statisics as we see fit but there are some type of statistics that its columns arent cacheable so their performances are really bad so i would like to know is there´s a way to add these columns to the cache to improve their performance? Thanks¡!¡I'm not aware of a concept of "cacheable" columns, at LEAST not in postgres itself; I've found a few references to higher level TOOLS that can use postgres referring to columns being cacheable or not. If not those, Are you thinking of Indexes? There are a few types that aren't indexable (Like Set types) but you can create Indexes fairly easily for any other type:

Code: [Select]CREATE INDEX testtabletable_testcolumn_index
    ON public.testtable USING btree
    (testcolumn)
    TABLESPACE pg_default;
You can also index via expressions. So if you wanted to index a set based on the RESULT of a stored procedure operating on that set, you COULD do something like:

Code: [Select]CREATE INDEX testtabletable_testcolumn_index
    ON public.testtable USING btree
    (EvaluateSet(testcolumn))
    TABLESPACE pg_default;
Where EvaluateSet is the function accepting a set type.