Keeping all datastreams in a single table/partition is not scalable as the number of datastream increases, thus we need to separate datastreams. Two possible options: - Create a different table for each datastream (a table named after the datastream) - Continue using the obs table but partition it by datastream and then by phenomenon time
Keeping all datastreams in a single table/partition is not scalable as the number of datastream increases, thus we need to separate datastreams.
Two possible options:
Create a different table for each datastream (a table named after the datastream)
Continue using the obs table but partition it by datastream and then by phenomenon time