(iceberg) branch main updated: Spark 3.5: Add Parallelism Parameter Validation to AddFilesProcedure. (#12784)

[etudenhoefner]

This is an automated email from the ASF dual-hosted git repository.

etudenhoefner pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/iceberg.git


The following commit(s) were added to refs/heads/main by this push:
     new 04bb3eed34 Spark 3.5: Add Parallelism Parameter Validation to 
AddFilesProcedure. (#12784)
04bb3eed34 is described below

commit 04bb3eed3451522a4ee53daf38b198853844d815
Author: slfan1989 <55643692+slfan1...@users.noreply.github.com>
AuthorDate: Tue Apr 22 14:57:19 2025 +0800

    Spark 3.5: Add Parallelism Parameter Validation to AddFilesProcedure. 
(#12784)
---
 .../iceberg/spark/extensions/TestAddFilesProcedure.java  | 16 ++++++++++++++++
 .../iceberg/spark/procedures/AddFilesProcedure.java      |  1 +
 2 files changed, 17 insertions(+)

diff --git 
a/spark/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestAddFilesProcedure.java
 
b/spark/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestAddFilesProcedure.java
index 56aa40aba9..141de608ea 100644
--- 
a/spark/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestAddFilesProcedure.java
+++ 
b/spark/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestAddFilesProcedure.java
@@ -1109,6 +1109,22 @@ public class TestAddFilesProcedure extends 
ExtensionsTestBase {
         sql("SELECT id, name, dept, subdept FROM %s ORDER BY id", tableName));
   }
 
+  @TestTemplate
+  public void testAddFilesWithInvalidParallelism() {
+    createUnpartitionedHiveTable();
+
+    createIcebergTable(
+        "id Integer, name String, dept String, subdept String", "PARTITIONED 
BY (id)");
+
+    assertThatThrownBy(
+            () ->
+                sql(
+                    "CALL %s.system.add_files(table => '%s', source_table => 
'%s', parallelism => -1)",
+                    catalogName, tableName, sourceTableName))
+        .isInstanceOf(IllegalArgumentException.class)
+        .hasMessage("Parallelism should be larger than 0");
+  }
+
   @TestTemplate
   public void testAddFilesToTableWithManySpecs() {
     createPartitionedHiveTable();
diff --git 
a/spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/procedures/AddFilesProcedure.java
 
b/spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/procedures/AddFilesProcedure.java
index 1e14b748c0..f2b3f8658f 100644
--- 
a/spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/procedures/AddFilesProcedure.java
+++ 
b/spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/procedures/AddFilesProcedure.java
@@ -118,6 +118,7 @@ class AddFilesProcedure extends BaseProcedure {
     boolean checkDuplicateFiles = input.asBoolean(CHECK_DUPLICATE_FILES_PARAM, 
true);
 
     int parallelism = input.asInt(PARALLELISM, 1);
+    Preconditions.checkArgument(parallelism > 0, "Parallelism should be larger 
than 0");
 
     return importToIceberg(
         tableIdent, sourceIdent, partitionFilter, checkDuplicateFiles, 
parallelism);