Hi Doris Devs,

For UNIQUE KEY tables, I see that partitioning is restricted to the key
columns.

In my case (multi-tenant system), the key is (AccountId, KeyId) where:

   -

   AccountId = tenant ID
   -

   KeyId = entity-specific ID
   Both have high cardinality. Partitioning by either can lead to tablet
   explosion,

Questions:

   1.

   Should I even partition such a table? My table size is in the range of *10
   MB to 400 MB*.
   2.

   I don’t need tenant-level data drops, so partition pruning isn’t a
   requirement.
   3.

   Without partitioning, how do I handle future data growth? With bucketing
   alone, a fixed bucket count must be chosen upfront, right?

Looking for guidance on the recommended industry standard strategy here.

Thanks,
~H

Reply via email to