SQL Support with JDBC about sidewinder HOT 5 CLOSED

I'll work on this.

from sidewinder.

I will start working on this issue

from sidewinder.

@scherukuri11 thanks

from sidewinder.

Sidewinder storage engines have been redesigned to follow true columnar storage design pattern with independent storage and compression for time and value columns. This has enabled the backend to support projection limitation on Apache Calcite query engine and Avatica server for direct JDBC. d7d5010

So far the proof of concept design works however a lot of corner cases have to be fixed and currently this is extremely brittle. Phase 2 of this development is to solidify the query planner and validate queries work as expected. Also extend Tag support to columns rather than single field and also add where clause to tag index mapping.

Phase 3 implementation will need to include memory optimization and tuning to ensure large queries are automatically capped so as to not run out of memory. Also additional tuning is needed for reducing memory footprint for high cardinality measurements. Investigation for pagination capabilities is also critical to reduce heap usage. Lastly implementation of calcite time operators (streaming) can be applied to time series functions in Sidewinder.

Note: that as of 0.2.x the key value pair based tag indices are now available allowing tags to be treated as traditional column values.

from sidewinder.

Committed phase 1 to development e4c713a

from sidewinder.