The backend of inverted index is done but still lack of GUI frontend for users.
We are doing PoC (proof of concept) of inverted index under the streaming architecture <https://github.com/apache/incubator-kylin/blob/master/docs/TechInside/Kylin%200.7%20Lambda%20Architecture.md> internally, and compare with Druid. So far the result is very promising. With OLAP cube as the historic storage, we expect much better query latency than Druid. There is no recent plan to support composite data types like array, map, structure. Our priority in the coming months is streaming and cube build performance improvement. Cheers Yang On Tue, Mar 31, 2015 at 11:37 AM, dong wang <[email protected]> wrote: > Hi, currently, is it available for us to use this feature to filter the raw > data without any aggregations? > > in addition, do we support complicated data types in HIVE such as > array, maps, structs? >
