GitHub user PragmaTwice edited a comment on the discussion: TimeSeries Proposal

> Personnaly I prefer some internal merging rules to "compact" them, which 
> requires a "dynamic" chunking

Is it too complicated to do compaction on chunks?

>From my side, it seems fixed-size (instead of fixed-time-range) chunking can 
>also work on Kvrocks? e.g.
1. chunk size is fixed instead of in a fixed time range;
2. chunk id is the timestamp of the first item in the chunk.

Then both range query and point query can be efficient (?)

And, for writing:
- appending: write to the HEAD chunk (create a new chunk if it's full)
- inserting (rare?): do a point query via chunk id, and write to the chunk 
(insert a new chunk if it's full)

WDYT? cc @mapleFU @yezhizi 




GitHub link: 
https://github.com/apache/kvrocks/discussions/3044#discussioncomment-13728089

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to