[GitHub] [iceberg] dimas-b commented on a diff in pull request #5037: Core: Implement BaseMetastoreCatalog.registerTable()

[GitBox <gi...@apache.org>]

dimas-b commented on code in PR #5037:
URL: https://github.com/apache/iceberg/pull/5037#discussion_r920179288


##########
nessie/src/test/java/org/apache/iceberg/nessie/TestNessieTable.java:
##########
@@ -385,6 +388,89 @@ public void testDropTable() throws IOException {
     verifyCommitMetadata();
   }
 
+  private void testRegister(TableIdentifier identifier, String metadataVersionFiles) {
+    Assertions.assertThat(catalog.registerTable(identifier, "file:" + metadataVersionFiles)).isNotNull();
+    Table newTable = catalog.loadTable(identifier);
+    Assertions.assertThat(newTable).isNotNull();
+    TableOperations ops = ((HasTableOperations) newTable).operations();
+    String metadataLocation = ((NessieTableOperations) ops).currentMetadataLocation();
+    Assertions.assertThat("file:" + metadataVersionFiles).isEqualTo(metadataLocation);
+    Assertions.assertThat(catalog.dropTable(identifier, false)).isTrue();
+  }
+
+  @Test
+  public void testRegisterTableWithGivenBranch() {
+    List<String> metadataVersionFiles = metadataVersionFiles(TABLE_NAME);
+    Assertions.assertThat(1).isEqualTo(metadataVersionFiles.size());
+    ImmutableTableReference tableReference =
+        ImmutableTableReference.builder().reference("main").name(TABLE_NAME).build();
+    TableIdentifier identifier = TableIdentifier.of(DB_NAME, tableReference.toString());
+    testRegister(identifier, metadataVersionFiles.get(0));
+  }
+
+  @Test
+  public void testRegisterTableNegativeScenarios() throws NessieConflictException, NessieNotFoundException {

Review Comment:
   "Negative" in what sense? Would you mind renaming to `testRegisterTableFailureScenarios`? Also, it would be preferable to have a separate test method for each case, IMHO.



##########
aws/src/integration/java/org/apache/iceberg/aws/dynamodb/TestDynamoDbCatalog.java:
##########
@@ -295,6 +298,37 @@ public void testDropNamespace() {
     Assert.assertFalse("namespace must not exist", response.hasItem());
   }
 
+  @Test
+  public void testRegisterTable() {
+    Namespace namespace = Namespace.of(genRandomName());

Review Comment:
   What is the purpose of using random names here? What do we achieve by randomization? Using random test input makes it harder to reproduce and debug failures :thinking: 



##########
nessie/src/test/java/org/apache/iceberg/nessie/TestNessieTable.java:
##########
@@ -385,6 +388,89 @@ public void testDropTable() throws IOException {
     verifyCommitMetadata();
   }
 
+  private void testRegister(TableIdentifier identifier, String metadataVersionFiles) {

Review Comment:
   This is a utility method, but it is names like a test method... Would you mind renaming to something like `validateRegister`?



##########
nessie/src/test/java/org/apache/iceberg/nessie/TestNessieTable.java:
##########
@@ -385,6 +388,89 @@ public void testDropTable() throws IOException {
     verifyCommitMetadata();
   }
 
+  private void testRegister(TableIdentifier identifier, String metadataVersionFiles) {
+    Assertions.assertThat(catalog.registerTable(identifier, "file:" + metadataVersionFiles)).isNotNull();
+    Table newTable = catalog.loadTable(identifier);
+    Assertions.assertThat(newTable).isNotNull();
+    TableOperations ops = ((HasTableOperations) newTable).operations();
+    String metadataLocation = ((NessieTableOperations) ops).currentMetadataLocation();
+    Assertions.assertThat("file:" + metadataVersionFiles).isEqualTo(metadataLocation);
+    Assertions.assertThat(catalog.dropTable(identifier, false)).isTrue();
+  }
+
+  @Test
+  public void testRegisterTableWithGivenBranch() {
+    List<String> metadataVersionFiles = metadataVersionFiles(TABLE_NAME);
+    Assertions.assertThat(1).isEqualTo(metadataVersionFiles.size());
+    ImmutableTableReference tableReference =
+        ImmutableTableReference.builder().reference("main").name(TABLE_NAME).build();
+    TableIdentifier identifier = TableIdentifier.of(DB_NAME, tableReference.toString());
+    testRegister(identifier, metadataVersionFiles.get(0));
+  }
+
+  @Test
+  public void testRegisterTableNegativeScenarios() throws NessieConflictException, NessieNotFoundException {
+    List<String> metadataVersionFiles = metadataVersionFiles(TABLE_NAME);
+    Assertions.assertThat(1).isEqualTo(metadataVersionFiles.size());
+    // Case 1: Branch does not exist
+    ImmutableTableReference defaultTableReference =
+        ImmutableTableReference.builder().reference("default").name(TABLE_NAME).build();
+    TableIdentifier defaultIdentifier = TableIdentifier.of(DB_NAME, defaultTableReference.toString());
+    Assertions.assertThatThrownBy(
+        () -> catalog.registerTable(
+            defaultIdentifier, "file:" + metadataVersionFiles.get(0)))
+        .isInstanceOf(IllegalArgumentException.class)
+        .hasMessage("Nessie ref 'default' does not exist");
+    // Case 2: Table Already Exists
+    Assertions.assertThatThrownBy(() -> catalog.registerTable(TABLE_IDENTIFIER, "file:" + metadataVersionFiles.get(0)))
+        .isInstanceOf(AlreadyExistsException.class)
+        .hasMessage("Table already exists: db.tbl");
+    // Case 3: Registering using a tag
+    ImmutableTableReference branchTableReference =
+        ImmutableTableReference.builder().reference(BRANCH).name(TABLE_NAME).build();
+    TableIdentifier branchIdentifier = TableIdentifier.of(DB_NAME, branchTableReference.toString());
+    Assertions.assertThat(catalog.dropTable(branchIdentifier, false)).isTrue();
+    String hash = api.getReference().refName(BRANCH).get().getHash();
+    api.createReference().sourceRefName(BRANCH).reference(Tag.of("tag_1", hash)).create();
+    ImmutableTableReference tagTableReference =
+        ImmutableTableReference.builder().reference("tag_1").name(TABLE_NAME).build();
+    TableIdentifier tagIdentifier = TableIdentifier.of(DB_NAME, tagTableReference.toString());
+    Assertions.assertThatThrownBy(
+        () -> catalog.registerTable(
+            tagIdentifier, "file:" + metadataVersionFiles.get(0)))
+        .isInstanceOf(IllegalArgumentException.class)
+        .hasMessage("You can only mutate tables when using a branch without a hash or timestamp.");
+    // Case 4: non-null metadata path with null metadata location
+    Assertions.assertThatThrownBy(
+        () -> catalog.registerTable(TABLE_IDENTIFIER, "file:" + metadataVersionFiles.get(0) + "invalidName"))
+        .isInstanceOf(NotFoundException.class);
+    // Case 5: null identifier
+    Assertions.assertThatThrownBy(
+        () -> catalog.registerTable(null, "file:" + metadataVersionFiles.get(0) + "invalidName"))
+        .isInstanceOf(IllegalArgumentException.class)
+        .hasMessage("Invalid identifier: null");
+  }
+
+  @Test
+  public void testRegisterTableWithDefaultBranch() {
+    List<String> metadataVersionFiles = metadataVersionFiles(TABLE_NAME);
+    Assertions.assertThat(1).isEqualTo(metadataVersionFiles.size());
+    Assertions.assertThat(catalog.dropTable(TABLE_IDENTIFIER, false)).isTrue();
+    testRegister(TABLE_IDENTIFIER, metadataVersionFiles.get(0));
+  }
+
+  @Test
+  public void testRegisterTableMoreThanOneBranch() {

Review Comment:
   Where are other branches in this test? I can only see operations on `main` :thinking: 



##########
nessie/src/main/java/org/apache/iceberg/nessie/NessieTableOperations.java:
##########
@@ -152,7 +152,8 @@ protected void doCommit(TableMetadata base, TableMetadata metadata) {
       }
     }
 
-    String newMetadataLocation = writeNewMetadata(metadata, currentVersion() + 1);
+    String newMetadataLocation = (base == null) && (metadata.metadataFileLocation() != null) ?
+        metadata.metadataFileLocation() : writeNewMetadata(metadata, currentVersion() + 1);

Review Comment:
   It looks like we're alternating between a simple get and a more involved write operation in this statement... Would you mind moving the write under an `if` to emphasize that it's a substantial operation?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org