mbutrovich commented on code in PR #5032:
URL: https://github.com/apache/datafusion-comet/pull/5032#discussion_r3716928583
##########
spark/src/test/resources/sql-tests/expressions/csv/to_csv.sql:
##########
Review Comment:
Now that `to_csv` routes through the dispatcher by default for a struct
built from `named_struct`, would it be worth adding a case where one of the
struct's fields is itself a cast that can raise under ANSI, for example
`to_csv(named_struct('a', CAST(bad_str AS INT)))`? #5219 found and fixed a bug
in the shared dispatcher for exactly this shape, and it merged after this
branch's last commit, so there is not yet a test here that pins the interaction
for `to_csv`.
Would it also be worth adding a query that runs `to_csv` over a struct with
an array, map, or nested-struct field, even without asserting the exact output
value? The comment block above explains well why Spark's own output for complex
types is not a stable value to assert, but that is a different question from
whether Comet's dispatcher path can hand a nested
`InternalRow`/`ArrayData`/`MapData` through `CometSpecializedGettersDispatch`
without hitting an unhandled exception of its own. That path is not exercised
by any test in this PR, and it is the same class of gap #5219 found for TIME
types.
--
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]
