You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@iceberg.apache.org by "hililiwei (via GitHub)" <gi...@apache.org> on 2023/04/04 08:00:46 UTC
[GitHub] [iceberg] hililiwei commented on a diff in pull request #7218: Flink: Use starting sequence number by default when rewriting data files
hililiwei commented on code in PR #7218:
URL: https://github.com/apache/iceberg/pull/7218#discussion_r1156834157
##########
core/src/main/java/org/apache/iceberg/actions/BaseRewriteDataFilesAction.java:
##########
@@ -198,6 +200,11 @@ public BaseRewriteDataFilesAction<ThisT> filter(Expression expr) {
return this;
}
+ public BaseRewriteDataFilesAction<ThisT> useStartingSequenceNumber(boolean useStarting) {
Review Comment:
add doc.
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
+ icebergTableUnPartitioned
+ .updateProperties()
+ .set(TableProperties.UPSERT_ENABLED, "true")
+ .commit();
+
+ // Add 1 data file and 1 equality-delete file
+ sql("INSERT INTO %s SELECT 1, 'hi'", TABLE_NAME_UNPARTITIONED);
+
+ icebergTableUnPartitioned.refresh();
+ CloseableIterable<FileScanTask> tasks = icebergTableUnPartitioned.newScan().planFiles();
+ List<DataFile> dataFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::file));
+ Set<DeleteFile> deleteFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::deletes)).stream()
+ .flatMap(Collection::stream)
+ .collect(Collectors.toSet());
+ Assert.assertEquals("Should have 3 data files before rewrite", 3, dataFiles.size());
+ Assert.assertTrue(
+ "Should have 1 equality-delete file before rewrite",
+ deleteFiles.size() == 1
+ && Iterables.getOnlyElement(deleteFiles).content() == FileContent.EQUALITY_DELETES);
+
+ Assert.assertThrows(
+ "Rewrite using new sequence number should fail",
+ ValidationException.class,
+ () ->
+ Actions.forTable(stale1).rewriteDataFiles().useStartingSequenceNumber(false).execute());
+ // Rewrite using the starting sequence number should succeed
+ RewriteDataFilesActionResult result = Actions.forTable(stale2).rewriteDataFiles().execute();
+
+ // Should not rewrite files from the new commit
+ Assert.assertEquals("Action should rewrite 2 data files", 2, result.deletedDataFiles().size());
+ Assert.assertEquals("Action should add 1 data file", 1, result.addedDataFiles().size());
+
+ icebergTableUnPartitioned.refresh();
+
+ // Assert the table records as expected.
+ SimpleDataUtil.assertTableRecords(
+ icebergTableUnPartitioned,
+ Lists.newArrayList(
+ SimpleDataUtil.createRecord(1, "hi"), SimpleDataUtil.createRecord(2, "world")));
+ }
+
+ private void upgradeToVersion2(BaseTable table) {
Review Comment:
We don't need to purposely create a method for this.
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
Review Comment:
Create a table with primary keys in before().
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
+ icebergTableUnPartitioned
+ .updateProperties()
+ .set(TableProperties.UPSERT_ENABLED, "true")
+ .commit();
+
+ // Add 1 data file and 1 equality-delete file
+ sql("INSERT INTO %s SELECT 1, 'hi'", TABLE_NAME_UNPARTITIONED);
+
+ icebergTableUnPartitioned.refresh();
+ CloseableIterable<FileScanTask> tasks = icebergTableUnPartitioned.newScan().planFiles();
+ List<DataFile> dataFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::file));
+ Set<DeleteFile> deleteFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::deletes)).stream()
+ .flatMap(Collection::stream)
+ .collect(Collectors.toSet());
+ Assert.assertEquals("Should have 3 data files before rewrite", 3, dataFiles.size());
+ Assert.assertTrue(
+ "Should have 1 equality-delete file before rewrite",
+ deleteFiles.size() == 1
+ && Iterables.getOnlyElement(deleteFiles).content() == FileContent.EQUALITY_DELETES);
+
+ Assert.assertThrows(
+ "Rewrite using new sequence number should fail",
+ ValidationException.class,
+ () ->
+ Actions.forTable(stale1).rewriteDataFiles().useStartingSequenceNumber(false).execute());
+ // Rewrite using the starting sequence number should succeed
+ RewriteDataFilesActionResult result = Actions.forTable(stale2).rewriteDataFiles().execute();
+
+ // Should not rewrite files from the new commit
+ Assert.assertEquals("Action should rewrite 2 data files", 2, result.deletedDataFiles().size());
+ Assert.assertEquals("Action should add 1 data file", 1, result.addedDataFiles().size());
+
+ icebergTableUnPartitioned.refresh();
Review Comment:
It doesn't need to.
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
+ icebergTableUnPartitioned
+ .updateProperties()
+ .set(TableProperties.UPSERT_ENABLED, "true")
+ .commit();
+
+ // Add 1 data file and 1 equality-delete file
+ sql("INSERT INTO %s SELECT 1, 'hi'", TABLE_NAME_UNPARTITIONED);
+
+ icebergTableUnPartitioned.refresh();
+ CloseableIterable<FileScanTask> tasks = icebergTableUnPartitioned.newScan().planFiles();
+ List<DataFile> dataFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::file));
+ Set<DeleteFile> deleteFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::deletes)).stream()
+ .flatMap(Collection::stream)
+ .collect(Collectors.toSet());
+ Assert.assertEquals("Should have 3 data files before rewrite", 3, dataFiles.size());
+ Assert.assertTrue(
+ "Should have 1 equality-delete file before rewrite",
+ deleteFiles.size() == 1
+ && Iterables.getOnlyElement(deleteFiles).content() == FileContent.EQUALITY_DELETES);
Review Comment:
Take it apart in two `Assert`
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
+ icebergTableUnPartitioned
+ .updateProperties()
+ .set(TableProperties.UPSERT_ENABLED, "true")
+ .commit();
+
+ // Add 1 data file and 1 equality-delete file
+ sql("INSERT INTO %s SELECT 1, 'hi'", TABLE_NAME_UNPARTITIONED);
+
+ icebergTableUnPartitioned.refresh();
+ CloseableIterable<FileScanTask> tasks = icebergTableUnPartitioned.newScan().planFiles();
+ List<DataFile> dataFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::file));
+ Set<DeleteFile> deleteFiles =
+ Lists.newArrayList(CloseableIterable.transform(tasks, FileScanTask::deletes)).stream()
+ .flatMap(Collection::stream)
+ .collect(Collectors.toSet());
+ Assert.assertEquals("Should have 3 data files before rewrite", 3, dataFiles.size());
+ Assert.assertTrue(
+ "Should have 1 equality-delete file before rewrite",
+ deleteFiles.size() == 1
+ && Iterables.getOnlyElement(deleteFiles).content() == FileContent.EQUALITY_DELETES);
+
+ Assert.assertThrows(
Review Comment:
` Assertions#assertThatThrownBy` is recommended.
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
Review Comment:
Create a table with primary keys in before().
##########
flink/v1.16/flink/src/test/java/org/apache/iceberg/flink/actions/TestRewriteDataFilesAction.java:
##########
@@ -386,4 +395,73 @@ public void testRewriteAvoidRepeateCompress() throws IOException {
expected.add(SimpleDataUtil.createRecord(2, "b"));
SimpleDataUtil.assertTableRecords(icebergTableUnPartitioned, expected);
}
+
+ @Test
+ public void testRewriteNoConflictWithEqualityDeletes() throws IOException {
+ // Add 2 data files
+ sql("INSERT INTO %s SELECT 1, 'hello'", TABLE_NAME_UNPARTITIONED);
+ sql("INSERT INTO %s SELECT 2, 'world'", TABLE_NAME_UNPARTITIONED);
+
+ // Load 2 stale tables to pass to rewrite actions
+ Table stale1 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+ Table stale2 =
+ validationCatalog.loadTable(TableIdentifier.of(icebergNamespace, TABLE_NAME_UNPARTITIONED));
+
+ // Enabled upsert
+ icebergTableUnPartitioned.refresh();
+ upgradeToVersion2((BaseTable) icebergTableUnPartitioned);
+ icebergTableUnPartitioned
+ .updateSchema()
+ .allowIncompatibleChanges()
+ .requireColumn("id")
+ .setIdentifierFields("id")
+ .commit();
+ icebergTableUnPartitioned
+ .updateProperties()
+ .set(TableProperties.UPSERT_ENABLED, "true")
+ .commit();
+
+ // Add 1 data file and 1 equality-delete file
+ sql("INSERT INTO %s SELECT 1, 'hi'", TABLE_NAME_UNPARTITIONED);
+
+ icebergTableUnPartitioned.refresh();
Review Comment:
It doesn't need to.
--
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