From 1bbcae19a042a1285c0a06fb3b0997efd3144d88 Mon Sep 17 00:00:00 2001
From: Neil Conway <neil.conway@gmail.com>
Date: Sat, 19 Sep 2026 13:17:29 -0400
Subject: [PATCH] fix: handle null lists with empty children in Spark shuffle

---
 .../spark/src/function/array/shuffle.rs       | 40 ++++++++++++++++++-
 .../test_files/spark/array/shuffle.slt        |  6 +++
 2 files changed, 44 insertions(+), 2 deletions(-)

diff --git a/datafusion/spark/src/function/array/shuffle.rs b/datafusion/spark/src/function/array/shuffle.rs
index 2673c9155fe08..3b9d91f16d17f 100644
--- a/datafusion/spark/src/function/array/shuffle.rs
+++ b/datafusion/spark/src/function/array/shuffle.rs
@@ -184,8 +184,7 @@ fn general_array_shuffle<O: OffsetSizeTrait>(
         // skip the null value
         if array.is_null(row_index) {
             nulls.push(false);
-            offsets.push(offsets[row_index] + O::one());
-            mutable.try_extend(0, 0, 1)?;
+            offsets.push(offsets[row_index]);
             continue;
         }
         nulls.push(true);
@@ -269,9 +268,46 @@ fn fixed_size_array_shuffle(
 #[cfg(test)]
 mod tests {
     use super::*;
+    use arrow::array::{Int32Array, ListArray};
+    use arrow::buffer::NullBuffer;
     use arrow::datatypes::Field;
     use datafusion_expr::ReturnFieldArgs;
 
+    #[test]
+    fn test_shuffle_null_lists() -> Result<()> {
+        let field = Arc::new(Field::new_list_field(DataType::Int32, false));
+        let input = ListArray::new(
+            Arc::clone(&field),
+            OffsetBuffer::new(vec![0, 0, 0, 1, 1].into()),
+            Arc::new(Int32Array::from(vec![42])),
+            Some(NullBuffer::from(vec![false, true, true, false])),
+        );
+        let empty_child = ListArray::new(
+            Arc::clone(&field),
+            OffsetBuffer::new(vec![0, 0, 0].into()),
+            Arc::new(Int32Array::from(Vec::<i32>::new())),
+            Some(NullBuffer::from(vec![false, true])),
+        );
+
+        for (list, expected_values_len) in
+            [(empty_child, 0), (input.clone(), 1), (input.slice(3, 1), 0)]
+        {
+            for data_type in [list.data_type().clone(), LargeList(Arc::clone(&field))] {
+                let array = arrow::compute::cast(&list, &data_type)?;
+                let result = array_shuffle_with_seed(&[Arc::clone(&array)], Some(0))?;
+                assert_eq!(result.as_ref(), array.as_ref());
+                // Null rows need no placeholder child value.
+                let values_len = match data_type {
+                    List(_) => as_list_array(&result)?.values().len(),
+                    LargeList(_) => as_large_list_array(&result)?.values().len(),
+                    _ => unreachable!(),
+                };
+                assert_eq!(values_len, expected_values_len);
+            }
+        }
+        Ok(())
+    }
+
     #[test]
     fn test_shuffle_nullability() {
         let shuffle = SparkShuffle::new();
diff --git a/datafusion/sqllogictest/test_files/spark/array/shuffle.slt b/datafusion/sqllogictest/test_files/spark/array/shuffle.slt
index 061d89c4fac66..b0acee18fd225 100644
--- a/datafusion/sqllogictest/test_files/spark/array/shuffle.slt
+++ b/datafusion/sqllogictest/test_files/spark/array/shuffle.slt
@@ -45,6 +45,12 @@ SELECT shuffle(NULL);
 ----
 NULL
 
+# Typed null lists have no child values to shuffle.
+query ??
+SELECT shuffle(arrow_cast(NULL, 'List(Int32)')), shuffle(arrow_cast(NULL, 'LargeList(Int32)'));
+----
+NULL NULL
+
 # Test shuffle function with fixed size list arrays
 query ?
 SELECT shuffle(arrow_cast([1, 2, NULL, 3, 4, 5], 'FixedSizeList(6, Int64)'), 1);