fix: substring with negative indices should produce correct result (#470

)
apache · May 26, 2024 · 7ba5693 · 7ba5693
1 parent 5bfe46d
commit 7ba5693
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/core/src/execution/datafusion/planner.rs b/core/src/execution/datafusion/planner.rs
@@ -395,7 +395,8 @@ impl PhysicalPlanner {
                 let child = self.create_expr(expr.child.as_ref().unwrap(), input_schema)?;
                 // Spark Substring's start is 1-based when start > 0
                 let start = expr.start - i32::from(expr.start > 0);
-                let len = expr.len;
+                // substring negative len is treated as 0 in Spark
+                let len = std::cmp::max(expr.len, 0);
 
                 Ok(Arc::new(SubstringExec::new(
                     child,

diff --git a/spark/src/test/scala/org/apache/comet/CometExpressionSuite.scala b/spark/src/test/scala/org/apache/comet/CometExpressionSuite.scala
@@ -156,6 +156,12 @@ class CometExpressionSuite extends CometTestBase with AdaptiveSparkPlanHelper {
   test("string type and substring") {
     withParquetTable((0 until 5).map(i => (i.toString, (i + 100).toString)), "tbl") {
       checkSparkAnswerAndOperator("SELECT _1, substring(_2, 2, 2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT _1, substring(_2, 2, -2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT _1, substring(_2, -2, 2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT _1, substring(_2, -2, -2) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT _1, substring(_2, -2, 10) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT _1, substring(_2, 0, 0) FROM tbl")
+      checkSparkAnswerAndOperator("SELECT _1, substring(_2, 1, 0) FROM tbl")
     }
   }