[SPARK-44873] Support alter view with nested columns in Hive client

### What changes were proposed in this pull request?
Previously, if a view's schema contains a nested struct, alterTable using Hive client will fail. This change supports a view with a nested struct. The mechanism is to store an empty schema when we call Hive client, since we already store the actual schema in table properties. This fix is similar to apache#37364

### Why are the changes needed?
This supports using view with nested structs in Hive metastore.

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
Added unit test.

Closes apache#42532 from kylerong-db/hive_view.

Authored-by: kylerong-db <kyle.rong@databricks.com>
Signed-off-by: Gengliang Wang <gengliang@apache.org>

sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveExternalCatalog.scala

@@ -595,7 +595,16 @@ private[spark] class HiveExternalCatalog(conf: SparkConf, hadoopConf: Configurat
 
     if (tableDefinition.tableType == VIEW) {
       val newTableProps = tableDefinition.properties ++ tableMetaToTableProps(tableDefinition).toMap
-      client.alterTable(tableDefinition.copy(properties = newTableProps))
+      val newTable = tableDefinition.copy(properties = newTableProps)
+      try {
+        client.alterTable(newTable)
+      } catch {
+        case NonFatal(e) =>
+          // If for some reason we fail to store the schema we store it as empty there
+          // since we already store the real schema in the table properties. This try-catch
+          // should only be necessary for Spark views which are incompatible with Hive
+          client.alterTable(newTable.copy(schema = EMPTY_DATA_SCHEMA))
+      }
     } else {
       val oldTableDef = getRawTable(db, tableDefinition.identifier.table)
 

sql/hive/src/test/scala/org/apache/spark/sql/hive/HiveParquetSourceSuite.scala

@@ -385,4 +385,14 @@ class HiveParquetSourceSuite extends ParquetPartitioningTest with ParquetTest {
       checkAnswer(spark.table("t"), Row(Row("a", 1)))
     }
   }
+
+  test("Alter view with nested struct") {
+    withView("t", "t2") {
+      sql("CREATE OR REPLACE VIEW t AS SELECT " +
+        "struct(id AS `$col2`, struct(id AS `$col`) AS s1) AS s2 FROM RANGE(5)")
+      sql("ALTER VIEW t SET TBLPROPERTIES ('x' = 'y')")
+      sql("ALTER VIEW t RENAME TO t2")
+      checkAnswer(sql("show TBLPROPERTIES t2 (x)"), Row("x", "y"))
+    }
+  }
 }

sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/HiveDDLSuite.scala

@@ -175,14 +175,6 @@ class HiveCatalogedDDLSuite extends DDLSuite with TestHiveSingleton with BeforeA
     withView("v") {
       spark.sql("CREATE VIEW v AS SELECT STRUCT('a' AS `a`, 1 AS b) q")
       checkAnswer(sql("SELECT q.`a`, q.b FROM v"), Row("a", 1) :: Nil)
-
-      checkError(
-        exception = intercept[SparkException] {
-          spark.sql("ALTER VIEW v AS SELECT STRUCT('a' AS `$a`, 1 AS b) q")
-        },
-        errorClass = "CANNOT_RECOGNIZE_HIVE_TYPE",
-        parameters = Map("fieldType" -> "\"STRUCT<$A:STRING,B:INT>\"", "fieldName" -> "`q`")
-      )
     }
   }