Add view support in GlueSync

CHANGELOG.md

@@ -3,6 +3,13 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/) and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## 8.1.13 - 2025-10-31
+### Changed
+- Upgrade maven shade plugin to `3.6.1` (was `3.1.1`).
+- Upgrade AWS Java SDK in `apiary-gluesync-listener` to `1.12.792` (was `1.12.276`)
+- Send view original texts in `apiary-gluesync-listener`.
+- Add posibility to only sync views in `apiary-gluesync-listener` CLI
+
 ## 8.1.12 - 2025-09-15
 ### Added
 - GlueSyncCli with fat jar for on-demand syncing.

hive-event-listeners/apiary-gluesync-listener/pom.xml

@@ -11,6 +11,11 @@
   <artifactId>apiary-gluesync-listener</artifactId>
   <name>Apiary GlueSync Listener</name>
   <description>Glue Sync Listener for Apiary that replays metadata events in AWS Glue catalog</description>
+
+  <properties>
+    <aws-java-sdk.version>1.12.792</aws-java-sdk.version>
+  </properties>
+
   <dependencies>
     <dependency>
       <groupId>com.expediagroup.apiary</groupId>
@@ -137,7 +142,7 @@
       <plugin>
         <groupId>org.apache.maven.plugins</groupId>
         <artifactId>maven-shade-plugin</artifactId>
-        <version>3.1.1</version>
+        <version>3.6.1</version>
         <executions>
           <execution>
             <id>shade-all</id>

hive-event-listeners/apiary-gluesync-listener/src/main/java/com/expediagroup/apiary/extensions/gluesync/cli/GlueSyncCli.java

@@ -47,6 +47,7 @@
 import com.hotels.hcommon.hive.metastore.iterator.PartitionIterator;
 
 public class GlueSyncCli {
+
   private static final Logger logger = LoggerFactory.getLogger(GlueSyncCli.class);
 
   private static final String THRIFT_CONNECTION_URI = System.getenv("THRIFT_CONNECTION_URI");
@@ -108,9 +109,10 @@ public void syncAll(CommandLine cmd) throws TException {
 
     boolean continueOnError = cmd.hasOption("continueOnError");
     boolean deleteGluePartitions = !cmd.hasOption("keep-glue-partitions");
+    boolean syncOnlyViews = cmd.hasOption("sync-only-views");
 
-    logger.debug("Additional parameters: continueOnError={}, deleteGluePartitions={}", continueOnError,
-        deleteGluePartitions);
+    logger.debug("Additional parameters: continueOnError={}, deleteGluePartitions={}, syncOnlyViews={}",
+        continueOnError, deleteGluePartitions, syncOnlyViews);
 
     boolean hadError = false;
     for (String dbName : metastoreClient.getAllDatabases()) {
@@ -120,7 +122,7 @@ public void syncAll(CommandLine cmd) throws TException {
           if (tableName.matches(tableRegex)) {
             try {
               logger.info("Syncing table: {} in database: {}", tableName, dbName);
-              syncTable(dbName, tableName, deleteGluePartitions, verbose);
+              syncTable(dbName, tableName, deleteGluePartitions, syncOnlyViews, verbose);
             } catch (Exception e) {
               hadError = true;
               logger.error("Error syncing table: {} in database: {}: {}", tableName, dbName, e.getMessage());
@@ -139,8 +141,8 @@ public void syncAll(CommandLine cmd) throws TException {
     }
   }
 
-  private void syncTable(String dbName, String tableName, boolean deleteGluePartitions, boolean verbose)
-      throws TException {
+  private void syncTable(String dbName, String tableName, boolean deleteGluePartitions, boolean syncOnlyViews,
+      boolean verbose) throws TException {
     Database database = metastoreClient.getDatabase(dbName);
 
     if (!glueDatabaseService.exists(database)) {
@@ -150,6 +152,14 @@ private void syncTable(String dbName, String tableName, boolean deleteGluePartit
 
     Table table = metastoreClient.getTable(dbName, tableName);
 
+    if (syncOnlyViews) {
+      String type = table.getTableType();
+      if (type != null && !type.equalsIgnoreCase("VIRTUAL_VIEW")) {
+        logger.info("Table {}.{} is not a view, skipping as syncOnlyViews flag is active", dbName, tableName);
+        return;
+      }
+    }
+
     CreateTableEvent createTableEvent = new CreateTableEvent(table, true, null);
     apiaryGlueSync.onCreateTable(createTableEvent);
 
@@ -176,5 +186,4 @@ protected Iterator<Partition> createPartitionIterator(IMetaStoreClient metastore
       throws org.apache.hadoop.hive.metastore.api.MetaException, org.apache.thrift.TException {
     return new PartitionIterator(metastoreClient, table, DEFAULT_PARTITION_BATCH_SIZE);
   }
-
 }

hive-event-listeners/apiary-gluesync-listener/src/main/java/com/expediagroup/apiary/extensions/gluesync/cli/GlueSyncCliParser.java

@@ -108,6 +108,8 @@ private static CommandLineParser getParser(Options options) {
     options.addOption(new Option("c", "continueOnError", false, "Continue on error (default: false)"));
     options.addOption(new Option(null, "keep-glue-partitions", false,
         "If true, will keep glue partitions even if there is no corresponding hive partition. If false will delete them (default: false)"));
+    options.addOption(new Option(null, "sync-only-views", false,
+        "If true, it will only sync VIRTUAL_VIEW table type (default: false)"));
 
     CommandLineParser parser = new DefaultParser();
     return parser;

hive-event-listeners/apiary-gluesync-listener/src/main/java/com/expediagroup/apiary/extensions/gluesync/listener/service/HiveToGlueTransformer.java

@@ -101,7 +101,9 @@ public TableInput transformTable(final Table table) {
         .withPartitionKeys(partitionKeys)
         .withRetention(table.getRetention())
         .withStorageDescriptor(sd)
-        .withTableType(table.getTableType());
+        .withTableType(table.getTableType())
+        .withViewOriginalText(table.getViewOriginalText())
+        .withViewExpandedText(table.getViewExpandedText());
   }
 
   public PartitionInput transformPartition(final Partition partition) {

hive-event-listeners/apiary-gluesync-listener/src/test/java/com/expediagroup/apiary/extensions/gluesync/cli/GlueSyncCliTest.java

@@ -133,6 +133,7 @@ private Options createOptions() {
     options.addOption(new Option("h", "help", false, "Print usage information"));
     options.addOption(new Option(null, "continueOnError", false, "Continue processing on errors"));
     options.addOption(new Option(null, "keep-glue-partitions", false, "Keep existing Glue partitions"));
+    options.addOption(new Option(null, "sync-only-views", false, "If true, it will only sync VIRTUAL_VIEW table type"));
 
     return options;
   }
@@ -170,7 +171,6 @@ public void testSyncAllWithMatchingDatabaseAndTable() throws Exception {
     verify(mockApiaryGlueSync, never()).onCreateDatabase(any(CreateDatabaseEvent.class));
     verify(mockApiaryGlueSync, times(4)).onCreateTable(any(CreateTableEvent.class));
     verify(mockGluePartitionService, times(4)).synchronizePartitions(any(), any(), anyBoolean(), anyBoolean());
-
   }
 
   @Test
@@ -209,7 +209,6 @@ public void testSyncAllWithVerboseOption() throws Exception {
     verify(mockApiaryGlueSync, never()).onCreateDatabase(any(CreateDatabaseEvent.class));
     verify(mockApiaryGlueSync, times(1)).onCreateTable(any(CreateTableEvent.class));
     verify(mockGluePartitionService, times(1)).synchronizePartitions(any(), any(), anyBoolean(), eq(true));
-
   }
 
   @Test
@@ -262,7 +261,6 @@ public void testSyncAllWithNoMatchingDatabases() throws Exception {
     // Assert
     verify(mockApiaryGlueSync, never()).onCreateDatabase(any(CreateDatabaseEvent.class));
     verify(mockApiaryGlueSync, never()).onCreateTable(any(CreateTableEvent.class));
-
   }
 
   @Test(expected = RuntimeException.class)
@@ -339,7 +337,7 @@ public void testSyncAllWithKeepGluePartitionsOption() throws Exception {
     // Arrange - Test that deleteGluePartitions is false when keep-glue-partitions
     // is set
     String[] args = { "--database-name-regex", "test_db.*", "--table-name-regex", "test_table.*",
-        "--keep-glue-partitions" };
+                      "--keep-glue-partitions" };
     CommandLine cmd = new DefaultParser().parse(createOptions(), args);
 
     List<String> databases = Arrays.asList("test_db1");
@@ -369,6 +367,70 @@ public void testSyncAllWithKeepGluePartitionsOption() throws Exception {
     verify(mockGluePartitionService, times(1)).synchronizePartitions(any(), any(), eq(false), anyBoolean());
   }
 
+  @Test
+  public void testSyncAllWithSyncOnlyViewsOption_onView() throws Exception {
+    String[] args = { "--database-name-regex", "test_db.*", "--table-name-regex", "test_table.*",
+                      "--sync-only-views" };
+    CommandLine cmd = new DefaultParser().parse(createOptions(), args);
+
+    List<String> databases = Arrays.asList("test_db1");
+    List<String> tables = Arrays.asList("test_table1");
+
+    Database mockDatabase = new Database();
+    mockDatabase.setName("test_db1");
+
+    Table mockTable = new Table();
+    mockTable.setDbName("test_db1");
+    mockTable.setTableName("test_table1");
+    mockTable.setTableType("VIRTUAL_VIEW");
+    Map<String, String> tableParams = new HashMap<>();
+    mockTable.setParameters(tableParams);
+
+    when(mockMetastoreClient.getAllDatabases()).thenReturn(databases);
+    when(mockMetastoreClient.getAllTables("test_db1")).thenReturn(tables);
+    when(mockMetastoreClient.getDatabase(anyString())).thenReturn(mockDatabase);
+    when(mockMetastoreClient.getTable(anyString(), anyString())).thenReturn(mockTable);
+    when(mockIsIcebergTablePredicate.test(any())).thenReturn(false);
+    when(mockMetastoreClient.listPartitions(anyString(), anyString(), anyShort())).thenReturn(Arrays.asList());
+
+    // Act
+    glueSyncCli.syncAll(cmd);
+
+    // Assert - Verify that synchronizePartitions is called with
+    // deleteGluePartitions=false
+    verify(mockGluePartitionService, times(1)).synchronizePartitions(any(), any(), anyBoolean(), anyBoolean());
+  }
+
+  @Test
+  public void testSyncAllWithSyncOnlyViewsOption_onNormalTable() throws Exception {
+    String[] args = { "--database-name-regex", "test_db.*", "--table-name-regex", "test_table.*",
+                      "--sync-only-views" };
+    CommandLine cmd = new DefaultParser().parse(createOptions(), args);
+
+    List<String> databases = Arrays.asList("test_db1");
+    List<String> tables = Arrays.asList("test_table1");
+
+    Database mockDatabase = new Database();
+    mockDatabase.setName("test_db1");
+
+    Table mockTable = new Table();
+    mockTable.setDbName("test_db1");
+    mockTable.setTableName("test_table1");
+    mockTable.setTableType("EXTERNAL_TABLE");
+    Map<String, String> tableParams = new HashMap<>();
+    mockTable.setParameters(tableParams);
+
+    when(mockMetastoreClient.getAllDatabases()).thenReturn(databases);
+    when(mockMetastoreClient.getAllTables("test_db1")).thenReturn(tables);
+    when(mockMetastoreClient.getDatabase(anyString())).thenReturn(mockDatabase);
+    when(mockMetastoreClient.getTable(anyString(), anyString())).thenReturn(mockTable);
+
+    // Act
+    glueSyncCli.syncAll(cmd);
+
+    verify(mockGluePartitionService, times(0)).synchronizePartitions(any(), any(), anyBoolean(), anyBoolean());
+  }
+
   @Test
   public void testSyncAllWithDefaultPartitionDeletion() throws Exception {
     // Arrange - Test that deleteGluePartitions is true by default (when
@@ -443,7 +505,6 @@ public void testSyncAllWithLargePartitionBatchHandling() throws Exception {
     // Assert
     verify(mockGluePartitionService, times(1)).synchronizePartitions(any(), eq(largePartitionList), anyBoolean(),
         eq(true));
-
   }
 
   @Test(expected = RuntimeException.class)
@@ -498,7 +559,6 @@ public void testSyncAllWithNonMatchingTableRegex() throws Exception {
     verify(mockApiaryGlueSync, never()).onCreateDatabase(any(CreateDatabaseEvent.class));
     verify(mockApiaryGlueSync, never()).onCreateTable(any(CreateTableEvent.class));
     verify(mockGluePartitionService, never()).synchronizePartitions(any(), any(), anyBoolean(), anyBoolean());
-
   }
 
   @Test
@@ -516,7 +576,6 @@ public void testSyncAllWithEmptyDatabaseList() throws Exception {
     verify(mockApiaryGlueSync, never()).onCreateDatabase(any(CreateDatabaseEvent.class));
     verify(mockApiaryGlueSync, never()).onCreateTable(any(CreateTableEvent.class));
     verify(mockGluePartitionService, never()).synchronizePartitions(any(), any(), anyBoolean(), anyBoolean());
-
   }
 
   @Test

hive-event-listeners/apiary-gluesync-listener/src/test/java/com/expediagroup/apiary/extensions/gluesync/listener/ApiaryGlueSyncTest.java

@@ -290,6 +290,30 @@ public void onCreateIcebergTable() throws MetaException {
     assertThat(toList(createTableRequest.getTableInput().getStorageDescriptor().getColumns()), is(asList(colNames)));
   }
 
+  @Test
+  public void onCreateHiveView() throws MetaException {
+    CreateTableEvent event = mock(CreateTableEvent.class);
+    when(event.getStatus()).thenReturn(true);
+
+    Table table = simpleHiveTable(simpleSchema(), Collections.emptyList());
+    table.setTableType("VIRTUAL_VIEW");
+    table.setViewOriginalText("SELECT * FROM some_table");
+    table.setViewExpandedText("SELECT * FROM some_table");
+    when(event.getTable()).thenReturn(table);
+
+    glueSync.onCreateTable(event);
+
+    verify(glueClient).createTable(createTableRequestCaptor.capture());
+    verify(metricService).incrementCounter(MetricConstants.LISTENER_TABLE_SUCCESS);
+    CreateTableRequest createTableRequest = createTableRequestCaptor.getValue();
+
+    assertThat(createTableRequest.getDatabaseName(), is(gluePrefix + dbName));
+    assertThat(createTableRequest.getTableInput().getName(), is(tableName));
+    assertThat(createTableRequest.getTableInput().getTableType(), is("VIRTUAL_VIEW"));
+    assertThat(createTableRequest.getTableInput().getViewOriginalText().isEmpty(), is(false));
+    assertThat(createTableRequest.getTableInput().getViewExpandedText().isEmpty(), is(false));
+  }
+
   @Test
   public void onAlterHiveTable() throws MetaException {
     AlterTableEvent event = mock(AlterTableEvent.class);