GitHub user PragmaTwice edited a comment on the discussion: TimeSeries Proposal
Yeah, for optimizing the write performance, we should limit IO as less as we can. So it will be a huge speedup if we don't modify the metadata in every chunk write. `first_timestamp` and `last_timestamp` is easy to get via a one-step prefix iteration in subkey CF, and also as @mapleFU mentioned, `TS.INFO` is not a high-frequently-called commands. But, it seems even we remove these fields, there are also `size`, `chunk_size`.. in the metadata. GitHub link: https://github.com/apache/kvrocks/discussions/3044#discussioncomment-13726976 ---- This is an automatically sent email for issues@kvrocks.apache.org. To unsubscribe, please send an email to: issues-unsubscr...@kvrocks.apache.org
