Fix: array contains null handling

spark/src/main/scala/org/apache/comet/serde/arrays.scala

@@ -22,6 +22,7 @@ package org.apache.comet.serde
 import scala.annotation.tailrec
 
 import org.apache.spark.sql.catalyst.expressions.{ArrayAppend, ArrayContains, ArrayDistinct, ArrayExcept, ArrayFilter, ArrayInsert, ArrayIntersect, ArrayJoin, ArrayMax, ArrayMin, ArrayRemove, ArrayRepeat, ArraysOverlap, ArrayUnion, Attribute, CreateArray, ElementAt, Expression, Flatten, GetArrayItem, IsNotNull, Literal, Reverse, Size}
+import org.apache.spark.sql.catalyst.util.GenericArrayData
 import org.apache.spark.sql.internal.SQLConf
 import org.apache.spark.sql.types._
 
@@ -134,7 +135,34 @@ object CometArrayContains extends CometExpressionSerde[ArrayContains] {
 
     val arrayContainsScalarExpr =
       scalarFunctionExprToProto("array_has", arrayExprProto, keyExprProto)
-    optExprWithInfo(arrayContainsScalarExpr, expr, expr.children: _*)
+
+    // Handle NULL array input - return NULL if array is NULL (matching Spark's behavior)
+    val isNotNullExpr = createUnaryExpr(
+      expr,
+      expr.children.head,
+      inputs,
+      binding,
+      (builder, unaryExpr) => builder.setIsNotNull(unaryExpr))
+
+    val nullLiteralProto = exprToProto(Literal(null, BooleanType), Seq.empty)
+
+    if (arrayContainsScalarExpr.isDefined && isNotNullExpr.isDefined &&
+      nullLiteralProto.isDefined) {
+      val caseWhenExpr = ExprOuterClass.CaseWhen
+        .newBuilder()
+        .addWhen(isNotNullExpr.get)
+        .addThen(arrayContainsScalarExpr.get)
+        .setElseExpr(nullLiteralProto.get)
+        .build()
+      Some(
+        ExprOuterClass.Expr
+          .newBuilder()
+          .setCaseWhen(caseWhenExpr)
+          .build())
+    } else {
+      withInfo(expr, expr.children: _*)
+      None
+    }
   }
 }
 
@@ -395,6 +423,15 @@ object CometCreateArray extends CometExpressionSerde[CreateArray] {
       inputs: Seq[Attribute],
       binding: Boolean): Option[ExprOuterClass.Expr] = {
     val children = expr.children
+
+    // Handle empty array: return literal directly to avoid DataFusion coerce_types bug
+    // when make_array is called with 0 arguments (issue #3338)
+    if (children.isEmpty) {
+      val emptyArrayLiteral =
+        Literal.create(new GenericArrayData(Array.empty[Any]), expr.dataType)
+      return exprToProtoInternal(emptyArrayLiteral, inputs, binding)
+    }
+
     val childExprs = children.map(exprToProtoInternal(_, inputs, binding))
 
     if (childExprs.forall(_.isDefined)) {

spark/src/test/resources/sql-tests/expressions/array/array_contains.sql

@@ -35,5 +35,24 @@ query spark_answer_only
 SELECT array_contains(array(1, 2, 3), val) FROM test_array_contains
 
 -- literal + literal
-query ignore(https://github.com/apache/datafusion-comet/issues/3345)
+-- Note: array_contains(array(), 1) still has a bug (issue #3346) so we use spark_answer_only
+-- The NULL array case (cast(NULL as array<int>)) was fixed in issue #3345
+query spark_answer_only
 SELECT array_contains(array(1, 2, 3), 2), array_contains(array(1, 2, 3), 4), array_contains(array(), 1), array_contains(cast(NULL as array<int>), 1)
+
+-- Additional NULL array tests (issue #3345 fix verification)
+-- NULL array with integer value
+query
+SELECT array_contains(cast(NULL as array<int>), 1)
+
+-- NULL array with string value
+query
+SELECT array_contains(cast(NULL as array<string>), 'test')
+
+-- NULL array with NULL value
+query
+SELECT array_contains(cast(NULL as array<int>), cast(NULL as int))
+
+-- NULL array with column value
+query
+SELECT array_contains(cast(NULL as array<int>), val) FROM test_array_contains

spark/src/test/scala/org/apache/comet/CometArrayExpressionSuite.scala

@@ -325,6 +325,38 @@ class CometArrayExpressionSuite extends CometTestBase with AdaptiveSparkPlanHelp
     }
   }
 
+  test("array_contains - NULL array returns NULL") {
+    // Test that array_contains returns NULL when the array argument is NULL
+    // This matches Spark's SQL three-valued logic behavior
+    withTempDir { dir =>
+      withTempView("t1") {
+        val path = new Path(dir.toURI.toString, "test.parquet")
+        makeParquetFileAllPrimitiveTypes(path, dictionaryEnabled = false, n = 100)
+        spark.read.parquet(path.toString).createOrReplaceTempView("t1")
+
+        // Test NULL array with non-null value
+        checkSparkAnswerAndOperator(
+          sql("SELECT array_contains(cast(null as array<int>), 1) FROM t1"))
+        checkSparkAnswerAndOperator(
+          sql("SELECT array_contains(cast(null as array<string>), 'test') FROM t1"))
+        checkSparkAnswerAndOperator(
+          sql("SELECT array_contains(cast(null as array<double>), 1.5) FROM t1"))
+
+        // Test NULL array with NULL value
+        checkSparkAnswerAndOperator(
+          sql("SELECT array_contains(cast(null as array<int>), cast(null as int)) FROM t1"))
+
+        // Test NULL array with column value
+        checkSparkAnswerAndOperator(
+          sql("SELECT array_contains(cast(null as array<int>), _2) FROM t1"))
+
+        // Test non-null array with values (to ensure fix doesn't break normal operation)
+        checkSparkAnswerAndOperator(sql("SELECT array_contains(array(1, 2, 3), 2) FROM t1"))
+        checkSparkAnswerAndOperator(sql("SELECT array_contains(array(1, 2, 3), 5) FROM t1"))
+      }
+    }
+  }
+
   test("array_contains - test all types (convert from Parquet)") {
     withTempDir { dir =>
       val path = new Path(dir.toURI.toString, "test.parquet")