Hi Shuyi, Julian should know better on the effort of standardizing streaming SQL into the standard.
Regarding the engineering effort: I am trying to drive the implementation in both Calcite and Beam. Current TUMBLE as a table-valued function is working in Calcite master [1] while HOP and SESSION are in a PR [2] that is pending for review. I raised an open discussion for EMIT syntax[3]. You can also check the umbrella jira of this effort[4]. [1]:https://github.com/apache/calcite/pull/1481, https://github.com/apache/calcite/pull/1587 [2]:https://github.com/apache/calcite/pull/1761 [3]: https://lists.apache.org/thread.html/r5bd9a6f7af2c0cd81aecd4de512fd889fbf15f112cc3704f188b1d4f%40%3Cdev.calcite.apache.org%3E [4]:https://jira.apache.org/jira/browse/CALCITE-3271 -Rui On Mon, Feb 10, 2020 at 1:14 PM Shuyi Chen <[email protected]> wrote: > Hi all, > > I am wondering what the current status of Streaming SQL standardization is. > Has the work presented at the "One SQL to Rule Them All" SIGMOD paper being > standardized as ANSI/INCITS SQL, or what the current status is? Are there > any plan in Calcite/Beam/Flink to support the standards if so? AFAIK, there > is a BEAM proposal > < > https://docs.google.com/document/d/138uA7VTpbF84CFrd--cz3YVe0-AQ9ALnsavaSE2JeE4/edit#heading=h.f7eonhc93kwo > > > in > the BEAM community to move towards the new Streaming SQL extension. Thanks > a lot. > > -- > Shuyi Chen > https://www.linkedin.com/in/shuyi-chen-pku/ >
