[D] Problem getting s2 cell Ids for certain geometries [sedona]

["kalyan57 (via GitHub)" <gi...@apache.org>]

GitHub user kalyan57 created a discussion: Problem getting s2 cell Ids for certain geometries

hi folks!

There is a geo (multipolygon, attached)

And I can't get its s2 cell Ids using sedona API call ST_S2CellIds(my_geo, 17). It just gets stuck, it freezes so that spark executors just stop responding heartbeat requests and  my application fails.

my env: sedona 1.5.3 with scala 2.12.10 and spark 3.0.1.

the geometry is valid ST_IsValid(my_geo) returns true.

ST_Dump returns set of polygons. And yeah, of course, looking at the visualization that almost vertical streight line looks like something not desirerd, but anyway the geometry itself is valid... So why can't I just get its s2 cellIds?

The problem may be solved with a slight buffer (0.0000001 is enough), but in order to avoid buffering when it's not necessary, but I can't even determine what's wrong with this particular geometry to filter it out and process with buffering.

I saw that close-to-zero-wigth polygon that looks like a line, and yeah, that really looks like something to get rid of. But how to find such cases algorithmically?  

GitHub link: https://github.com/apache/sedona/discussions/1988

----
This is an automatically sent email for issues@sedona.apache.org.
To unsubscribe, please send an email to: issues-unsubscribe@sedona.apache.org