DEV Community

Discussion on: find the partition key with a duplicate covering index local read

Collapse
 
denismagda profile image
Denis Magda YugabyteDB

It's coincidence but I was researching how to query a geo-partitioned cluster efficiently when all you have is an id of a record (with no geo-column). Very timely! Just curious how big would be an impact for writes as long as the app needs to update all the indexes across distant regions.

Btw, love the planetary data model!

Collapse
 
franckpachot profile image
Franck Pachot YugabyteDB

Maybe not a coincidence as this was driven by discussions with colleagues and users 😀
Only the inserts and delete have to update all indexes because I don't expect an update in the primary key. Except if you include more columns in the indexes -> I'll add a comment on this suggestion, yes, good point