I have a table with three fields & composite primary key:
- SensorID, primary
- Timestamp, primary
- Value
I want Postgres to partition on SensorID. I do not know ahead of time what the SensorIDs will be. Postgres will need to generate new partitions as new SensorIDs arrive.
Questions:
- Should I be using a hash or list based partition?
- Is it possible to achieve this behaviour without using triggers?