szehon-ho commented on PR #12667:
URL: https://github.com/apache/iceberg/pull/12667#issuecomment-3309081814

   I chat yesterday with @mkaravel to get back on this topic, unfortunately 
afaik we are stuck a little on waiting consensus on 
https://github.com/apache/iceberg/pull/13227  , looks like there was a meeting 
about it when i was on vacation?  
   
   If i understand, the goal is to have linear edge interpolation for 
wraparound crs.  The debate is between
   
   1. Supporting wraparound CRS's for Geometry type.
   2. Support Linear edges for Geography type
   
   For (1) i think Menelaos's concern was , now the code needs to understand 
CRS in order to perform intersect.  I think otherwise the intersect logic may 
be incorrect in those wraparound CRS cases?  Or if we assume the bounds are 
always written correctly, it makes engine either complex to implement writing 
bounds that work in both cases, or writing inefficient bounds?
   
   For (2) I think Dewey did not want this approach as its not common?
   
   My knowledge is not so deep so forgive me if some part here is incorrect.  
   
   Was brainstorming other approaches with Menelaos yesterday, maybe either to 
enhance the spec to either add another type , Geometry with Wraparound, or 
enhance the physical encoding to add a boolean for wraparound, are good 
workarounds.
   
   Maybe we need another meeting?  But I think if we are going to implement the 
bounds/intersect logic we need to get consensus on this issue somehow.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to