Re: SLOPE - Planner optimizations on monotonic expressions.

On Wed, Jul 1, 2026 at 11:00 PM Zsolt Parragi <zsolt.parragi@percona.com>
wrote:

> Apologies for the late response, somehow I missed this update.
>

Thank you for your late response :D


> The changes generally look good and seem to work, except one remaining
> issue with domains:


I didn't know about domains before, tempting
e.g. cos(pi * x) decreasing for x between 0 and 1.
But I will avoid that rabbit hole.

The solution I implemented was using getBaseType
./src/backend/utils/cache/lsyscache.c:2754:getBaseType(Oid typid)

+ if(var_type >= FirstNormalObjectId)
+ var_type = getBaseType(var_type);

 /*
>   * get_const_sign
>   *             Helper to determine the sign of a numeric constant.
>   *             Returns 1 for positive, -1 for negative, 0 for zero or
> unknown.
>   */
> -static int
> +static inline enum NUMERIC_SIGN
>
> This comment is now outdated. The default value is unreachable
> currently, but is 0 still a safe value there?
>

Updated, actually NUMERIC_SIGN_ZERO indicates that the value
is strictly equals to 0, I don't see how it would be unsafe.


> +enum NUMERIC_SIGN {
> +       NUMERIC_SIGN_NINF=-2,
> +       NUMERIC_SIGN_NEG=-1,
> +       NUMERIC_SIGN_ZERO=0,
> +       NUMERIC_SIGN_POS=1,
> +       NUMERIC_SIGN_PINF=3,
> +       NUMERIC_SIGN_NAN=4,
> +       NUMERIC_SIGN_NULL=5,
> +};
>
> this is missing a typedef and maybe it should be defined at the
> beginning of the file or in a header? Is skipping positive 2
> intentional?
>

OK, moved to the top with a typedef, the 2 gap was not intentional,
I think when I wrote it I had NaN there before +Infinity, but later
I realised that NaN goes after +Infinity.


> The patchset also has many small formatting issues/inconsistencies,
> maybe it would be worth to run pgindent on it?
>

I ran it on src/backend/utils/adt/misc.c

It kept changing
/*
 * arg0_asc_slope_support
 *      Prosupport: f(x, ...) is monotonically increasing in x.
 */
to
/*
 * arg0_asc_slope_support Prosupport: f(x, ...) is monotonically increasing
 * in x.
 */

Is there a written standard/recommendation for these, maybe
/*
 * arg0_asc_slope_support
 *
 * Prosupport: f(x, ...) is monotonically increasing in x.
 */

The last part of the slope test is checking 480 corner cases, in my machine
it takes 240ms while some other tests take 1000+ so I guess it is not a
problem.


Regards,
Alexandre