Opened 9 years ago
Last modified 9 years ago
#13361 closed enhancement
What is the bbox of an incomplete OSM primitive? — at Initial Version
| Reported by: | GerdP | Owned by: | team |
|---|---|---|---|
| Priority: | normal | Milestone: | 16.12 |
| Component: | Core | Version: | |
| Keywords: | Cc: | bastiK, simon04, michael2402 |
Description
I stumbled over this while analysing performance problems in the
QuadBuckets.search(BBox searchBbox) method.
I noticed that incomplete primitives have quite confusing bboxes:
1) an incomplete node has
xmin = xmax = ymin = ymax = 0;
2) an incomplete way has
xmin = Double.POSITIVE_INFINITY
xmax = Double.NEGATIVE_INFINITY
ymin = Double.POSITIVE_INFINITY
ymax = Double.NEGATIVE_INFINITY
if all nodes are incomplete, else it will
be the bbox of the complete node(s).
3) a relation with eg. just two node members
where one is complete at (10,40) and the other
is incomplete will span a huge area from
xmin = 0
xmax = 40
ymin = 0
ymax = 10
I would expect that the bbox of a relation is only
calculated using the complete nodes.
I'd also prefer when the bbox of incomplete nodes would not be located
at (0,0), but I assume that there are good reasons for this.


