This is an automated email from the ASF dual-hosted git repository.
github-merge-queue[bot] pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/datafusion.git
The following commit(s) were added to refs/heads/main by this push:
new 5313ae9be2 fix: Correct `array_repeat` NULL handling (#25230)
5313ae9be2 is described below
commit 5313ae9be2f66a95c4a43f0337e1ef5cd6ad6933
Author: Neil Conway <[email protected]>
AuthorDate: Sat Sep 12 21:36:45 2026 +0000
fix: Correct `array_repeat` NULL handling (#25230)
## Which issue does this PR close?
- Closes #25223.
## Rationale for this change
Arrow allows elements of a `ListArray` to be logically NULL (according
to the validity buffer), but to have non-empty offset ranges. Such
elements should be considered NULL.
`array_repeat` only considered the offsets and ignored the validity
buffer, which resulted in mishandling such NULL values.
## What changes are included in this PR?
* Fix `array_repeat` bug
* Add unit test
* Add SLT
## What is the testing strategy for this PR?
Existing tests pass; new tests added. We add both a unit test and an SLT
because the SLT depends on `nullif` creating such a NULL value, which
might change in the future.
## Are there any user-facing changes?
Yes, bug fixed.
---
datafusion/functions-nested/src/repeat.rs | 53 +++++++++++++++++++---
.../sqllogictest/test_files/array/array_repeat.slt | 32 +++++++++++++
2 files changed, 79 insertions(+), 6 deletions(-)
diff --git a/datafusion/functions-nested/src/repeat.rs
b/datafusion/functions-nested/src/repeat.rs
index d7dff21141..b275ae661e 100644
--- a/datafusion/functions-nested/src/repeat.rs
+++ b/datafusion/functions-nested/src/repeat.rs
@@ -262,8 +262,16 @@ fn general_list_repeat<O: OffsetSizeTrait>(
continue;
};
let list_is_valid = list_array.is_valid(row_idx);
- let start = list_offsets[row_idx].to_usize().unwrap();
- let end = list_offsets[row_idx + 1].to_usize().unwrap();
+ // A NULL list slot may still have a non-empty offset range; treat the
+ // range as empty to keep the inner offsets and taken values in step.
+ let (start, end) = if list_is_valid {
+ (
+ list_offsets[row_idx].to_usize().unwrap(),
+ list_offsets[row_idx + 1].to_usize().unwrap(),
+ )
+ } else {
+ (0, 0)
+ };
let row_len = end - start;
for _ in 0..count {
@@ -272,9 +280,7 @@ fn general_list_repeat<O: OffsetSizeTrait>(
let offset = checked_repeat_offset::<O>(inner_running)?;
inner_offsets.push(offset);
inner_nulls.append(list_is_valid);
- if list_is_valid {
- take_indices.extend(start as u64..end as u64);
- }
+ take_indices.extend(start as u64..end as u64);
}
}
@@ -388,7 +394,7 @@ mod tests {
use super::{array_repeat_inner, general_list_repeat, general_repeat};
use arrow::array::{Array, ArrayRef, AsArray, Int32Array, Int64Array,
ListArray};
use arrow::buffer::{NullBuffer, OffsetBuffer, ScalarBuffer};
- use arrow::datatypes::{Field, Int32Type};
+ use arrow::datatypes::{DataType, Field, Int32Type};
use datafusion_common::Result;
use std::sync::Arc;
@@ -445,6 +451,41 @@ mod tests {
Ok(())
}
+ #[test]
+ fn test_array_repeat_nested_null_list_with_nonempty_range() -> Result<()> {
+ // A NULL list slot might still have a non-empty offset range
+ // https://github.com/apache/datafusion/issues/25223
+ let values = Int32Array::from(vec![1, 2, 3, 4, 5, 6]);
+ let list_array = ListArray::new(
+ Arc::new(Field::new_list_field(DataType::Int32, true)),
+ OffsetBuffer::new(ScalarBuffer::from(vec![0, 3, 6])),
+ Arc::new(values),
+ Some(NullBuffer::from(vec![false, true])),
+ );
+ let counts = Int64Array::from(vec![2, 2]);
+
+ let result = general_list_repeat::<i32>(&list_array, &counts)?;
+ let repeated_values = ListArray::from_iter_primitive::<Int32Type, _,
_>(vec![
+ None,
+ None,
+ Some(vec![Some(4), Some(5), Some(6)]),
+ Some(vec![Some(4), Some(5), Some(6)]),
+ ]);
+ let expected = ListArray::new(
+ Arc::new(Field::new_list_field(
+ repeated_values.data_type().clone(),
+ true,
+ )),
+ OffsetBuffer::new(ScalarBuffer::from(vec![0, 2, 4])),
+ Arc::new(repeated_values),
+ None,
+ );
+
+ assert_eq!(result.as_list::<i32>(), &expected);
+
+ Ok(())
+ }
+
#[test]
fn scalar_count_exceeding_max_array_size_returns_error() {
let element: ArrayRef = Arc::new(Int64Array::from(vec![1]));
diff --git a/datafusion/sqllogictest/test_files/array/array_repeat.slt
b/datafusion/sqllogictest/test_files/array/array_repeat.slt
index 5073c7d4c5..97e55e7718 100644
--- a/datafusion/sqllogictest/test_files/array/array_repeat.slt
+++ b/datafusion/sqllogictest/test_files/array/array_repeat.slt
@@ -238,5 +238,37 @@ NULL [NULL, NULL, NULL]
statement ok
drop table array_repeat_large_nested_null_table
+# https://github.com/apache/datafusion/issues/25223
+statement ok
+create table array_repeat_null_slot_table
+as values
+([1, 2, 3], [1, 2, 3], 1),
+([4, 5, 6], [7], 2);
+
+query I?
+select column3, array_repeat(nullif(column1, column2), 2)
+from array_repeat_null_slot_table order by column3;
+----
+1 [NULL, NULL]
+2 [[4, 5, 6], [4, 5, 6]]
+
+query I?
+select
+ column3,
+ array_repeat(
+ nullif(
+ arrow_cast(column1, 'LargeList(Int64)'),
+ arrow_cast(column2, 'LargeList(Int64)')
+ ),
+ 2
+ )
+from array_repeat_null_slot_table order by column3;
+----
+1 [NULL, NULL]
+2 [[4, 5, 6], [4, 5, 6]]
+
+statement ok
+drop table array_repeat_null_slot_table
+
include ./cleanup.slt.part
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
