N-Boutaib commented on PR #20646:
URL: https://github.com/apache/datafusion/pull/20646#issuecomment-3997717497

   > 👋 Hi! If I am understanding this PR correctly, then it sounds like it is 
accommodating substrait plans which are incorrect. Why are producers "widening" 
types in substrait plans?
   > 
   > Do we have producers that are doing this in practice or is it the DF 
producer itself which is producing these plans?
   > 
   > If it is a problem with the plan itself, then I think we should reject the 
plan altogether.
   > 
   > Please let me know if I am misunderstanding :)
   
   👋 The plans aren't incorrect — the widening happens because our logical 
Substrait producer has a simplified type system that normalizes all integer 
variants to 64-bit. This is intentional; it doesn't carry storage-level 
precision. When DataFusion resolves the table via NamedTable, the catalog 
reports the actual storage type (Int32), and the strict check rejects it even 
though the promotion is lossless.


-- 
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