Re: view canonicalization - looking for database gurus to chime in

[Jiang Xingbo]

Hi all,

I have recently prepared a design document for Spark SQL robust view
canonicalization, in the doc we defined the expected behavior and described
a late binding approach, views created by older versions of Spark/HIVE are
still supposed to work under this new approach.
For more details, please review:
https://docs.google.com/document/d/16NEDA9ejzAQAkm_tRqEVpaUAXgXuUo5gTQsbnItOUsA

This doc is accomplished under the guidance of Herman van Hovell, I've
learned a lot from the process.
Thanks to Reynold Xin, Srinath Shankar for the valuable advices and for
helping improve the quality of this design doc.
Any suggestions and discussions are welcomed, thank you all!

Sincerely,
Xingbo



--
View this message in context: 
http://apache-spark-developers-list.1001551.n3.nabble.com/view-canonicalization-looking-for-database-gurus-to-chime-in-tp19681p19996.html
Sent from the Apache Spark Developers List mailing list archive at Nabble.com.

-
To unsubscribe e-mail: dev-unsubscr...@spark.apache.org

view canonicalization - looking for database gurus to chime in

[Reynold Xin]

I know there are a lot of people with experience on developing database
internals on this list. Please take a look at this proposal for a new,
simpler way to handle view canonicalization in Spark SQL:
https://issues.apache.org/jira/browse/SPARK-18209

It sounds much simpler than what we currently do in 2.0/2.1, but I'm not
sure if there are obvious holes that I missed.