Magnus, I recommend that in addition to the mailing list that you drop in on the Drill weekly hangout so that you can ask your questions interactively. That can make a huge difference for questions where you don't quite know how to ask the question.
On Thu, May 28, 2015 at 1:47 AM, Magnus Pierre <[email protected]> wrote: > Hello dev, > > Please see https://issues.apache.org/jira/browse/DRILL-3180 < > https://issues.apache.org/jira/browse/DRILL-3180>, I have developed a > quick and dirty JDBC storage-plugin for Apache Drill. > Any help appreciated. > Next step is to replace JDBC classes with Calcite classes. On a local copy > I’ve already done this for the connection piece and now show tables works > in sqlline which is good. > > Next step is to get push down working and make it simpler by using the > great stuff in Calcite. > > I’m considering how the integration with Calcite will work in the > mentioned plugin. What I would like to do is to apply all the nice rules in > Calcite and then generate the target SQL from that, based on the current > plan and plugin. I’m not 100% w regards to the plans in Drill so any > pointers to material on that would be a great read. > > I’ve seen in the JDBC adapter code for Calcite that Calcite can go from > logical plan to SQL (I just had a quick read through so I might have > misunderstood something), any ideas, pointers or anything that can help me > figure how Calcite and Drill works together and where it would fit in would > be great. > > Regards, > Magnus > >
