Wondering if anyone has attempted to implement collection partitioning within MongoDb? I have some time-series-ish data which I’d like to partition so that I can simply drop stale partitions.
In my case I could have 30 to 60 partitions (days) and I would either need to query every partition for certain criteria (when most likely it’s just the latest partition which contains data I care about), or, I’d have to build some smarts to track which collections contain data I need to process.
Is there an aggregation framework operation which can span multiple collections (containing the same document structure)?
I could use application-level joins/aggregation but I suspect the performance wouldn’t be good if the database was hosted on a separate box.
Thanks for any thoughts/advice.