[jira] [Created] (CALCITE-1546) Wrong result/plan for NOT IN subqueries with disjunction

2016-12-19 Thread Vineet Garg (JIRA)
Vineet Garg created CALCITE-1546: Summary: Wrong result/plan for NOT IN subqueries with disjunction Key: CALCITE-1546 URL: https://issues.apache.org/jira/browse/CALCITE-1546 Project: Calcite

Re: Pushing order by to source

2016-12-19 Thread Gupta, Shuchi
Hey Jesus Thanks for your response. Yes, the sorting for queries having time series or group by clause are pushed downstream. I face issues when the query only has order by(in which case, the queryType is 'select') .. somehow in the RelOptCluster I am able to get the sort order from original

Re: Pushing order by to source

2016-12-19 Thread Gian Merlino
I think the main difference between the two is that Druid's built-in SQL uses Druid classes internally to model things, rather than just sticking to the HTTP API like the Druid rules in Calcite. If you think it makes sense for Calcite's Druid adapter to depend on Druid jars to get those classes

Re: Towards Calcite release 1.11

2016-12-19 Thread Julian Hyde
So, I’m hearing that we should make a release soonish. I propose code freeze by Wed 28th (9 days from now), hopefully release in 1st or 2nd week of January. Does that sound reasonable? Does anyone volunteer to be a release manager? I’d rather not be release manager again, but I’ll do it if no

Re: Pushing order by to source

2016-12-19 Thread Julian Hyde
Gian, It’s off-topic for this thread, but I wanted to ask whether you have any plans to keep the Calcite rules in Druid in sync with the Druid rules in Calcite (the “druid adapter”). My goal is to enable two modes of operation: (a) native SQL for Druid, (b) hybrid systems that combine Druid

Re: Moderators

2016-12-19 Thread Josh Elser
Found out that apmail is apparently /dev/null these days. Finally filed https://issues.apache.org/jira/browse/INFRA-13134 and I should be a moderator henceforth. If you want to remove yourself, feel free to ask INFRA via JIRA, Ashutosh. Similarly, feel free to ask if you'd like to have

Re: Pushing order by to source

2016-12-19 Thread Gian Merlino
Fwiw, Druid has a built-in set of Calcite rules now, and it does support pushing down order by and limit. It can even re-order the grouping dimensions for better performance, and choose timeseries, topN, or groupBy as appropriate. Depending on what you're trying to do, this may work for you. The

RE: [NOTICE] - analysis.apache.org closing down 14th January

2016-12-19 Thread Anton Mushin
Hi, +1 for Sonar. I think it would to improve Calcite and required code's quality in future, also we can find new problems in code base possibly. Best regards, Anton Mushin -Original Message- From: Julian Hyde [mailto:jh...@apache.org] Sent: Sunday, December 18, 2016 4:08 AM To:

RE: Pushing order by to source

2016-12-19 Thread Gupta, Shuchi
I could see that if order by is present in the query, Calcite is handling this sorting on its own. Can you tell me the class where this sorting is happening. I tried to debug in SqlParserImpl class , in the OrderBy(boolean accept) method. But while debugging it appears this code is not being