[SPARK-54029][SQL] Add detailed error message for table metadata corruption to ease debugging

Author: ganeshas-db

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/catalog/interface.scala

@@ -451,7 +451,17 @@ case class CatalogTable(
    */
   def partitionSchema: StructType = {
     val partitionFields = schema.takeRight(partitionColumnNames.length)
-    assert(partitionFields.map(_.name) == partitionColumnNames)
+    val actualPartitionColumnNames = partitionFields.map(_.name)
+
+    assert(actualPartitionColumnNames == partitionColumnNames,
+      "Corrupted table metadata detected for table " + identifier.unquotedString + ". " +
+      "The partition column names in the table schema " +
+      "do not match the declared partition columns. " +
+      "Table schema columns: [" + schema.fieldNames.mkString(", ") + "] " +
+      "Declared partition columns: [" + partitionColumnNames.mkString(", ") + "] " +
+      "Actual partition columns from schema: " +
+      "[" + actualPartitionColumnNames.mkString(", ") + "]. " +
+      "This indicates corrupted table metadata that needs to be repaired.")
 
     StructType(partitionFields)
   }

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/catalog/SessionCatalogSuite.scala

@@ -1951,4 +1951,80 @@ abstract class SessionCatalogSuite extends AnalysisTest with Eventually {
       assert(catalog.getCachedTable(qualifiedName2) != null)
     }
   }
+
+  test("CatalogTable partitionSchema provides detailed error for corrupted metadata") {
+    // Test case 1: Partition columns don't match schema (wrong names)
+    val corruptedTable1 = CatalogTable(
+      identifier = TableIdentifier("corrupted_table1", Some("test_db")),
+      tableType = CatalogTableType.MANAGED,
+      storage = CatalogStorageFormat.empty,
+      schema = StructType(Seq(
+        StructField("id", IntegerType),
+        StructField("name", StringType),
+        StructField("year", IntegerType),
+        StructField("month", IntegerType)
+      )),
+      partitionColumnNames = Seq("year", "day")  // "day" doesn't exist in schema
+    )
+
+    val exception1 = intercept[AssertionError] {
+      corruptedTable1.partitionSchema
+    }
+
+    val expectedMessage1 = "assertion failed: Corrupted table metadata detected " +
+      "for table test_db.corrupted_table1. " +
+      "The partition column names in the table schema " +
+      "do not match the declared partition columns. " +
+      "Table schema columns: [id, name, year, month] " +
+      "Declared partition columns: [year, day] " +
+      "Actual partition columns from schema: [year, month]. " +
+      "This indicates corrupted table metadata that needs to be repaired."
+    assert(exception1.getMessage === expectedMessage1)
+
+    // Test case 2: Wrong number of partition columns
+    val corruptedTable2 = CatalogTable(
+      identifier = TableIdentifier("corrupted_table2", Some("test_db")),
+      tableType = CatalogTableType.MANAGED,
+      storage = CatalogStorageFormat.empty,
+      schema = StructType(Seq(
+        StructField("id", IntegerType),
+        StructField("data", StringType),
+        StructField("partition_col", StringType)
+      )),
+      partitionColumnNames = Seq("partition_col", "extra_col")  // Too many partition columns
+    )
+
+    val exception2 = intercept[AssertionError] {
+      corruptedTable2.partitionSchema
+    }
+
+    val expectedMessage2 = "assertion failed: Corrupted table metadata detected " +
+      "for table test_db.corrupted_table2. " +
+      "The partition column names in the table schema " +
+      "do not match the declared partition columns. " +
+      "Table schema columns: [id, data, partition_col] " +
+      "Declared partition columns: [partition_col, extra_col] " +
+      "Actual partition columns from schema: [data, partition_col]. " +
+      "This indicates corrupted table metadata that needs to be repaired."
+    assert(exception2.getMessage === expectedMessage2)
+
+    // Test case 3: Valid table should work without error
+    val validTable = CatalogTable(
+      identifier = TableIdentifier("valid_table", Some("test_db")),
+      tableType = CatalogTableType.MANAGED,
+      storage = CatalogStorageFormat.empty,
+      schema = StructType(Seq(
+        StructField("id", IntegerType),
+        StructField("name", StringType),
+        StructField("year", IntegerType),
+        StructField("month", IntegerType)
+      )),
+      partitionColumnNames = Seq("year", "month")  // Matches schema
+    )
+
+    // This should not throw an exception
+    val partitionSchema = validTable.partitionSchema
+    assert(partitionSchema.fieldNames.toSeq == Seq("year", "month"))
+    assert(partitionSchema.fields.length == 2)
+  }
 }