GROUP BY groups reversed segments into single bucket

[komzpa]

I'm making a traffic model of a city.

I have a table with (geom, speed), where geom is two-point line representing edge of a graph, and speed is a floating point value. Edges can be duplicate and can be reversed.

When I create summary, (select geom, avg(speed) from table group by 1), it groups reversed segments into same buckets.

Expected behavior: reversed segments go to different buckets.

[robe]

Your group by would be grouping by bounding box of the line string. Not following what you mean by reverse segments.

Perhaps you can provide a small example.

[robe]

Not enough information to do anything with. This I suspect is not an issue and just user misunderstanding of how group by works in regard to PostGIS.

[robe]

kompzpa says your commit for #3844 fixed this.