rebase

Author: coderfender

native/spark-expr/src/static_invoke/char_varchar_utils/read_side_padding.rs

@@ -100,14 +100,14 @@ fn spark_read_side_padding2(
                     truncate,
                     ColumnarValue::Scalar(ScalarValue::Int32(Some(*length))),
                     string,
-                    is_left_pad
+                    is_left_pad,
                 ),
                 DataType::LargeUtf8 => spark_read_side_padding_internal::<i64>(
                     array,
                     truncate,
                     ColumnarValue::Scalar(ScalarValue::Int32(Some(*length))),
                     string,
-                    is_left_pad
+                    is_left_pad,
                 ),
                 // Dictionary support required for SPARK-48498
                 DataType::Dictionary(_, value_type) => {
@@ -268,15 +268,13 @@ fn add_padding_string(
         } else {
             Ok(string)
         }
-    } else if is_left_pad {
-        println!("is lpad so flipping padding and the string");
-        let pad_needed = length - char_len;
-        let pad: String = pad_string.chars().cycle().take(pad_needed).collect();
-        Ok(format!("{}{}", pad, string))
     } else {
         let pad_needed = length - char_len;
         let pad: String = pad_string.chars().cycle().take(pad_needed).collect();
-        Ok(string + &pad)
-
+        if is_left_pad {
+            Ok(format!("{}{}", pad, string))
+        } else {
+            Ok(string + &pad)
+        }
     }
 }

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

@@ -189,16 +189,11 @@ object CometStringLPad extends CometExpressionSerde[StringLPad] {
       expr: StringLPad,
       inputs: Seq[Attribute],
       binding: Boolean): Option[Expr] = {
-    expr.pad match {
-      case Literal(str, DataTypes.StringType) if str.toString == " " =>
-        scalarFunctionExprToProto(
-          "lpad",
-          exprToProtoInternal(expr.str, inputs, binding),
-          exprToProtoInternal(expr.len, inputs, binding))
-      case _ =>
-        withInfo(expr, "StringLPad with non-space characters is not supported")
-        None
-    }
+    scalarFunctionExprToProto(
+      "lpad",
+      exprToProtoInternal(expr.str, inputs, binding),
+      exprToProtoInternal(expr.len, inputs, binding),
+      exprToProtoInternal(expr.pad, inputs, binding))
   }
 }
 

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

@@ -425,6 +425,24 @@ class CometExpressionSuite extends CometTestBase with AdaptiveSparkPlanHelper {
     }
   }
 
+  test("test lpad expression support") {
+    val data = Seq(("IfIWasARoadIWouldBeBent", 10), ("తెలుగు", 2))
+    withParquetTable(data, "t1") {
+      val res = sql("select lpad(_1,_2) , lpad(_1,2) from t1 order by _1")
+      checkSparkAnswerAndOperator(res)
+    }
+  }
+
+  test("LPAD with character support other than default space") {
+    val data = Seq(("IfIWasARoadIWouldBeBent", 10), ("hi", 2))
+    withParquetTable(data, "t1") {
+      val res = sql(
+        """ select lpad(_1,_2,'?'), lpad(_1,_2,'??') , lpad(_1,2, '??'), hex(lpad(unhex('aabb'), 5)),
+          rpad(_1, 5, '??') from t1 order by _1 """.stripMargin)
+      checkSparkAnswerAndOperator(res)
+    }
+  }
+
   test("dictionary arithmetic") {
     // TODO: test ANSI mode
     withSQLConf(SQLConf.ANSI_ENABLED.key -> "false", "parquet.enable.dictionary" -> "true") {