Hi Kumar, Currently elastic table row type is "Map<String, Object>". I'm working on converting elastic mapping directly to RelDataType [1]. When done, you'll be able to query document attributes directly (as in `select address from user` instead of `select _MAP['address'] from user`) (discussion here [2])
Regarding missing features in elastic adapter, we can discuss them here and create JIRAs individually. Please let me know what functionalities you need. If you would like to contribute to Calcite a good place to start is jira issues (https://issues.apache.org/jira/projects/CALCITE). Read tutorial ( https://calcite.apache.org/docs/tutorial.html) first and study CSV adapter code. Thanks, Andrei. [1] https://calcite.apache.org/apidocs/org/apache/calcite/rel/type/RelDataType.html [2] https://lists.apache.org/thread.html/4fdb9fb396730168704f6bd9d3b57a7baa5b41241f26db7c0cc84729@%3Cdev.calcite.apache.org%3E33[3]aaaa aa On Mon, Feb 18, 2019 at 6:15 AM Shashwat Kumar <shashwatkmr....@gmail.com> wrote: > Hi all, > > I find that calcite for elasticsearch is still at very early stage and does > not support even select queries seamlessly. For example, > *SELECT id, address FROM *user needs to be done as *SELECT _MAP['id'], > _MAP['address'] FROM 'user'* > > I am using calcite elasticsearch in my project and would like to contribute > in it. I would like to know if there is any priority list of items which I > can take or other ongoing developments in it. > Kindly guide me through other necessary resources or guidelines. > > -- > Regards > Shashwat Kumar >