AdamGS commented on PR #17759:
URL: https://github.com/apache/datafusion/pull/17759#issuecomment-3333112240

   After sleeping on it, my current thinking is:
   1. Most SQL oriented systems have on decimal type that hides the underlying 
size in some way or is just fixed size, I know there's work/thinking going into 
a more logical-type system in DataFusion, but for now introducing new types is 
a major change.
   2. Therefore, Decimal32/64 parsing from SQL should be behind some config 
(I'm not 100% sure if there's a precedence for that or how well wired 
configuration is to the parser - but that's a change that I can imagine being 
useful for other things in the future).
   3. The coercion logic should stay as it is here - trying to fit 
integers/floats into decimals when applicable, or the wider decimal when trying 
to operate on two decimal types.
   4. There are other changes that I ran into in this PR that could be split 
into a separate PR that is a natural followup to #17501 - 
`spark::width_bucket`, `abs`, SMJ support and creating zero scalars.


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


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

Reply via email to