sdf-jkl opened a new issue, #18419:
URL: https://github.com/apache/datafusion/issues/18419
### Describe the bug
While working on #18363 I saw an inconsistency within match arm logic for
different inner List types for `LargeList`.
```diff
-let (inner_field, inner_offsets, inner_values, nulls) = // in
LargeList(LargeList)
+let (inner_field, inner_offsets, inner_values, _) = // in LargeList(List)
as_list_array(&values)?.clone().into_parts();
let flattened_array = GenericListArray::<i64>::new(
inner_field,
offsets,
inner_values,
nulls, <------- the validity buffer is used here
);
```
### To Reproduce
To reproduce the error, I fixed an existing `sqllogictest` to have
`LargeList(nullable LargeList(nullable Int64))`
This way when we generate a `flattened_array` the validity buffer's len and
values difference will cause an error.
```
External error: task 26 panicked with message "called `Result::unwrap()` on
an `Err` value: InvalidArgumentError(\"Incorrect length of null buffer for
LargeListArray, expected 1 got 5\")"
```
### Expected behavior
The expected behavior is to pass the new `sqllogictest`
```
flatten(arrow_cast(make_array([1], null, [2, 3], [null], make_array(4, null,
5)), 'LargeList(LargeList(Int64))')),
----
[1, NULL, 2, 3, NULL, 4, NULL, 5]
```
### Additional context
_No response_
--
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]
