Like the TTLs in cassandra, providing a much better support for sliding time window caching / matching
Unfortunately there isn’t. The request is tracked here and is provisionally planned for the 2.2 release (3 releases from now).
Good to know. Thanks, Andrei.
Do we have a rough time line when 2.2 will be released?
In about a year. But the planning this far ahead is very speculative. If you particularly care about this feature, I’d encourage you to tell your story on that issue. User stories that draw sympathy influence our roadmap quite a bit
+1, this is a feature we would need much sooner than later. I believe I have commented on this in other forum posts and have seen a lot of support for it. I would suggest making this a priority for NoSQL user adoption.
1.2 has become 2.0. And the feature is now listed as 2.2. in our internal planning. But again, very speculative. No concrete plans.
I know in Cassandra row TTLs are implemented using Tombstones, where a row that has been deleted or TTL’d into non-existence is still a part of the data set until it’s removed via compaction (ETL merge process that combines SSTables). Until it’s removed it is still scanned as a part of queries, but excluded from the result set via the tombstone marker. Not exactly efficient for data sets that have random deletes, but it’s really nice for serial data sets like time series where you want to only keep data for certain time frames (which is my use case).
Any chance this is going to be looked at again any time soon?
Yes, it’s still outstanding.
The best way to follow this is with the GitHub issue you linked.