Hey all.
I’ve started working on an aggregation engine for Redisearch.
The internals are implemented and I have a clear vision on how it will work, even in distributed mode, but I haven’t written the API yet.
Here is a proposal for the user facing API. Of course it will be abstracted by client libraries, but I would still love some comments on the design.
https://gist.github.com/dvirsky/f3300b90bf943aefc48c5c5e6c658670
Thanks,
Dvir
tw-bert
2
Hi Dvir, great design.
About the terminology: I like what InfluxData has done (inspired by Graphite a.o.): https://docs.influxdata.com/influxdb/v1.3/query_language/functions/
So even when bucketing, there’s a distinction between aggregators, selectors, transformations, and predictors.
Possibly provides some idea’s for the final design.
Cheers, TW
dvirsky
3
Thanks, I’ll have a look at it.