This is an automated email from the ASF dual-hosted git repository.
viirya pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/arrow-rs.git
The following commit(s) were added to refs/heads/master by this push:
new c92943cee Fix list equal for empty offset list array (#1818)
c92943cee is described below
commit c92943cee1d93c234b4003c6ae5e4f68edc1e063
Author: Liang-Chi Hsieh <[email protected]>
AuthorDate: Thu Jun 9 09:17:24 2022 -0700
Fix list equal for empty offset list array (#1818)
* Fix list equal for empty offset list array
* For review
---
arrow/src/array/equal/list.rs | 5 +++++
arrow/src/array/equal/mod.rs | 51 +++++++++++++++++++++++++++++++++++++++++++
2 files changed, 56 insertions(+)
diff --git a/arrow/src/array/equal/list.rs b/arrow/src/array/equal/list.rs
index 65d320c00..0feefa7aa 100644
--- a/arrow/src/array/equal/list.rs
+++ b/arrow/src/array/equal/list.rs
@@ -73,6 +73,11 @@ pub(super) fn list_equal<T: OffsetSizeTrait>(
// however, one is more likely to slice into a list array and get a region
that has 0
// child values.
// The test that triggered this behaviour had [4, 4] as a slice of 1 value
slot.
+ // For the edge case that zero length list arrays are always equal.
+ if len == 0 {
+ return true;
+ }
+
let lhs_child_length = lhs_offsets[lhs_start + len].to_usize().unwrap()
- lhs_offsets[lhs_start].to_usize().unwrap();
diff --git a/arrow/src/array/equal/mod.rs b/arrow/src/array/equal/mod.rs
index b89a8fa53..c3b0bbc95 100644
--- a/arrow/src/array/equal/mod.rs
+++ b/arrow/src/array/equal/mod.rs
@@ -629,6 +629,57 @@ mod tests {
test_equal(&a, &b, false);
}
+ #[test]
+ fn test_empty_offsets_list_equal() {
+ let empty: Vec<i32> = vec![];
+ let values = Int32Array::from(empty);
+ let empty_offsets: [u8; 0] = [];
+
+ let a = ArrayDataBuilder::new(DataType::List(Box::new(Field::new(
+ "item",
+ DataType::Int32,
+ true,
+ ))))
+ .len(0)
+ .add_buffer(Buffer::from(&empty_offsets))
+ .add_child_data(values.data().clone())
+ .null_bit_buffer(Some(Buffer::from(&empty_offsets)))
+ .build()
+ .unwrap();
+
+ let b = ArrayDataBuilder::new(DataType::List(Box::new(Field::new(
+ "item",
+ DataType::Int32,
+ true,
+ ))))
+ .len(0)
+ .add_buffer(Buffer::from(&empty_offsets))
+ .add_child_data(values.data().clone())
+ .null_bit_buffer(Some(Buffer::from(&empty_offsets)))
+ .build()
+ .unwrap();
+
+ test_equal(&a, &b, true);
+
+ let c = ArrayDataBuilder::new(DataType::List(Box::new(Field::new(
+ "item",
+ DataType::Int32,
+ true,
+ ))))
+ .len(0)
+ .add_buffer(Buffer::from(vec![0i32, 2, 3, 4, 6, 7, 8].to_byte_slice()))
+ .add_child_data(
+ Int32Array::from(vec![1, 2, -1, -2, 3, 4, -3, -4])
+ .data()
+ .clone(),
+ )
+ .null_bit_buffer(Some(Buffer::from(vec![0b00001001])))
+ .build()
+ .unwrap();
+
+ test_equal(&a, &c, true);
+ }
+
// Test the case where null_count > 0
#[test]
fn test_list_null() {
