rowkey vs hash-based

Data Modeling
Enthusiast

rowkey vs hash-based

Hi,

I am at a loss for designing to choose, whether I should go for rowkey ....ie hash-based by partition or should I leave only hash-based. The reason, is that partition is taking extra 2 bytes. Query run

is a join of two tables and date partition is there besides PI. Or is there any better design?

Thanks and regards,

Raja