rdblue commented on issue #28026: [SPARK-31257][SQL] Unify create table syntax
URL: https://github.com/apache/spark/pull/28026#issuecomment-614372327
 
 
   @cloud-fan, the meaning of external is dictated by Hive because that's where 
it came from. Spark is trying to change Hive's meaning of external, while 
maintaining "compatibility", but you can't do both. Because we now have a 
catalog abstraction, we're hitting the problem directly: you can't delegate to 
a catalog _and_ dictate what makes sense in cases specific to that catalog.
   
   If you want to drop support for external, then get rid of the keyword. 
Otherwise, it needs to be passed so that Spark isn't making arbitrary choices 
about details specific to a catalog. To be clear, I support failing all create 
statements that use EXTERNAL. That's fine with me, but you can't parse it and 
not pass it along.
   
   For SERDEPROPERTIES and OPTIONS, I fail to see how it is a similar 
situation. In that case, Spark has introduced a situation where it has clauses 
mean the same thing, so I implemented them to be passed the same way. We can 
pass these using separate prefixes if you'd like and let the catalog 
implementations choose. The problem is that this distinction has zero value for 
end users.
   
   Spark is easier to use and better overall if these are synonyms, but I can't 
say I care that much about it. Let me know what prefix you would like to use 
for SERDEPROPERTIES.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to