Ordering between partitions.

[Aleksandar Stojadinovic <al...@nissatech.com>]

Hello,

I've been searching for this for some time and I don't know if I'm just 
not looking for the right thing or it is just something not to be done.

We are analyzing Cassandra as a time series storage for some sensory 
data uploaded from a mobile phone. Of course, we would use the wide row 
technique partitioned by some parameter. We are not sure should it be a 
(value type, rounded time stamp) pair or just a rounded time stamp or a 
programmatic concatenation of those two so we can take an advice about that.

But the main issue is how can we get values stored in different 
partitions in an ordered manner. Using a Java client it looks like it is 
not really possible. Should we order the values in code or select all 
the primary keys with an IN keyword (but it seems we lose the pagination 
option then)? Is there a common pattern for this situation? All in all, 
this one bugs me a lot.

Best regards,
Aleksandar