[VL] Add Spark array_sort support

apache · Jan 14, 2025 · 0294890 · 0294890
1 parent 5755368
commit 0294890
Show file tree

Hide file tree

Showing 3 changed files with 18 additions and 2 deletions.
diff --git a/...ends-velox/src/main/scala/org/apache/gluten/backendsapi/velox/VeloxSparkPlanExecApi.scala b/...ends-velox/src/main/scala/org/apache/gluten/backendsapi/velox/VeloxSparkPlanExecApi.scala
@@ -200,6 +200,10 @@ class VeloxSparkPlanExecApi extends SparkPlanExecApi {
     }
   }
 
+  override def genArraySortTransformer(substraitExprName: String, argument: ExpressionTransformer, function: ExpressionTransformer, expr: ArraySort): ExpressionTransformer = {
+    GenericExpressionTransformer(substraitExprName, Seq(argument, function), expr)
+  }
+
   /** Transform array exists to Substrait */
   override def genArrayExistsTransformer(
       substraitExprName: String,

diff --git a/backends-velox/src/test/scala/org/apache/gluten/execution/MiscOperatorSuite.scala b/backends-velox/src/test/scala/org/apache/gluten/execution/MiscOperatorSuite.scala
@@ -1166,6 +1166,18 @@ class MiscOperatorSuite extends VeloxWholeStageTransformerSuite with AdaptiveSpa
     }
   }
 
+  test("array_sort") {
+    withTable("t") {
+      sql("create table t (a array<string>) using parquet")
+      sql("insert into t values (array('a', 'acds', 'bcedf', 'dc'))")
+      runQueryAndCompare(
+        "select array_sort(a, (x, y) -> " +
+          "if(length(x) > length(y), 1, if(length(x) < length(y), -1, 0))) from t") {
+        checkGlutenOperatorMatch[ProjectExecTransformer]
+      }
+    }
+  }
+
   test("Support bool type filter in scan") {
     withTable("t") {
       sql("create table t (id int, b boolean) using parquet")

diff --git a/ep/build-velox/src/get_velox.sh b/ep/build-velox/src/get_velox.sh
@@ -16,8 +16,8 @@
 
 set -exu
 
-VELOX_REPO=https://github.com/oap-project/velox.git
-VELOX_BRANCH=2025_01_14
+VELOX_REPO=https://github.com/boneanxs/velox.git
+VELOX_BRANCH=array_sort
 VELOX_HOME=""
 
 OS=`uname -s`