You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@flink.apache.org by lz...@apache.org on 2020/06/11 02:33:03 UTC
[flink] branch release-1.11 updated: [FLINK-18232][hive] Fix Hive
streaming source bugs
This is an automated email from the ASF dual-hosted git repository.
lzljs3620320 pushed a commit to branch release-1.11
in repository https://gitbox.apache.org/repos/asf/flink.git
The following commit(s) were added to refs/heads/release-1.11 by this push:
new a80f769 [FLINK-18232][hive] Fix Hive streaming source bugs
a80f769 is described below
commit a80f7690367e9a07b19321f3b34d8fdba17295b8
Author: Jingsong Lee <ji...@gmail.com>
AuthorDate: Thu Jun 11 10:30:46 2020 +0800
[FLINK-18232][hive] Fix Hive streaming source bugs
This closes #12573
---
.../flink/connectors/hive/HiveTableSource.java | 7 +-
.../hive/read/DirectoryMonitorDiscovery.java | 32 +++++++-
.../hive/read/HiveMapredSplitReader.java | 49 +++++-------
.../hive/read/HiveTableFileInputFormat.java | 19 ++++-
.../connectors/hive/HiveTableSourceITCase.java | 10 ++-
.../hive/read/DirectoryMonitorDiscoveryTest.java | 90 ++++++++++++++++++++++
.../hive/read/HiveTableFileInputFormatTest.java | 50 ++++++++++++
7 files changed, 214 insertions(+), 43 deletions(-)
diff --git a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/HiveTableSource.java b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/HiveTableSource.java
index 080cea8..2be4443 100644
--- a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/HiveTableSource.java
+++ b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/HiveTableSource.java
@@ -46,6 +46,7 @@ import org.apache.flink.table.catalog.hive.client.HiveShimLoader;
import org.apache.flink.table.catalog.hive.descriptors.HiveCatalogValidator;
import org.apache.flink.table.catalog.hive.util.HiveReflectionUtils;
import org.apache.flink.table.data.RowData;
+import org.apache.flink.table.data.TimestampData;
import org.apache.flink.table.filesystem.FileSystemLookupFunction;
import org.apache.flink.table.filesystem.FileSystemOptions;
import org.apache.flink.table.functions.AsyncTableFunction;
@@ -87,7 +88,7 @@ import java.util.Map;
import java.util.Properties;
import java.util.stream.Collectors;
-import static org.apache.flink.table.filesystem.DefaultPartTimeExtractor.toMills;
+import static org.apache.flink.table.filesystem.DefaultPartTimeExtractor.toLocalDateTime;
import static org.apache.flink.table.filesystem.FileSystemOptions.PARTITION_TIME_EXTRACTOR_CLASS;
import static org.apache.flink.table.filesystem.FileSystemOptions.PARTITION_TIME_EXTRACTOR_KIND;
import static org.apache.flink.table.filesystem.FileSystemOptions.PARTITION_TIME_EXTRACTOR_TIMESTAMP_PATTERN;
@@ -283,7 +284,9 @@ public class HiveTableSource implements
}
String consumeOffset = configuration.get(STREAMING_SOURCE_CONSUME_START_OFFSET);
- long currentReadTime = toMills(consumeOffset);
+ // to Local zone mills instead of UTC mills
+ long currentReadTime = TimestampData.fromLocalDateTime(toLocalDateTime(consumeOffset))
+ .toTimestamp().getTime();
Duration monitorInterval = configuration.get(STREAMING_SOURCE_MONITOR_INTERVAL);
diff --git a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscovery.java b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscovery.java
index f027302..aa15c81 100644
--- a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscovery.java
+++ b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscovery.java
@@ -18,7 +18,9 @@
package org.apache.flink.connectors.hive.read;
+import org.apache.flink.annotation.VisibleForTesting;
import org.apache.flink.api.java.tuple.Tuple2;
+import org.apache.flink.table.data.TimestampData;
import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
@@ -26,6 +28,7 @@ import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hive.metastore.api.Partition;
import java.io.IOException;
+import java.sql.Timestamp;
import java.util.ArrayList;
import java.util.List;
@@ -41,18 +44,39 @@ public class DirectoryMonitorDiscovery implements PartitionDiscovery {
Context context, long previousTimestamp) throws Exception {
FileStatus[] statuses = getFileStatusRecurse(
context.tableLocation(), context.partitionKeys().size(), context.fileSystem());
+ List<Tuple2<List<String>, Long>> partValueList = suitablePartitions(context, previousTimestamp, statuses);
+
List<Tuple2<Partition, Long>> partitions = new ArrayList<>();
+ for (Tuple2<List<String>, Long> tuple2 : partValueList) {
+ context.getPartition(tuple2.f0).ifPresent(
+ partition -> partitions.add(new Tuple2<>(partition, tuple2.f1)));
+ }
+ return partitions;
+ }
+
+ /**
+ * Find suitable partitions, extract timestamp and compare it with previousTimestamp.
+ */
+ @VisibleForTesting
+ static List<Tuple2<List<String>, Long>> suitablePartitions(
+ Context context,
+ long previousTimestamp,
+ FileStatus[] statuses) {
+ List<Tuple2<List<String>, Long>> partValueList = new ArrayList<>();
for (FileStatus status : statuses) {
List<String> partValues = extractPartitionValues(
new org.apache.flink.core.fs.Path(status.getPath().toString()));
long timestamp = context.extractTimestamp(
- context.partitionKeys(), partValues, status::getModificationTime);
+ context.partitionKeys(),
+ partValues,
+ // to UTC millisecond.
+ () -> TimestampData.fromTimestamp(
+ new Timestamp(status.getModificationTime())).getMillisecond());
if (timestamp >= previousTimestamp) {
- context.getPartition(partValues).ifPresent(
- partition -> partitions.add(new Tuple2<>(partition, timestamp)));
+ partValueList.add(new Tuple2<>(partValues, timestamp));
}
}
- return partitions;
+ return partValueList;
}
private static FileStatus[] getFileStatusRecurse(Path path, int expectLevel, FileSystem fs) {
diff --git a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveMapredSplitReader.java b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveMapredSplitReader.java
index bc4ff13..843f746 100644
--- a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveMapredSplitReader.java
+++ b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveMapredSplitReader.java
@@ -74,18 +74,12 @@ public class HiveMapredSplitReader implements SplitReader {
//StructObjectInspector in hive helps us to look into the internal structure of a struct object.
private final StructObjectInspector structObjectInspector;
- // Remember whether a row instance is reused. No need to set partition fields for reused rows
- private boolean rowReused = false;
-
- //Necessary info to init deserializer
- private final List<String> partitionKeys;
-
private final DataFormatConverters.DataFormatConverter[] converters;
- private final HiveTablePartition hiveTablePartition;
-
private final HiveShim hiveShim;
+ private final GenericRowData row;
+
public HiveMapredSplitReader(
JobConf jobConf,
List<String> partitionKeys,
@@ -93,7 +87,7 @@ public class HiveMapredSplitReader implements SplitReader {
int[] selectedFields,
HiveTableInputSplit split,
HiveShim hiveShim) throws IOException {
- this.hiveTablePartition = split.getHiveTablePartition();
+ HiveTablePartition hiveTablePartition = split.getHiveTablePartition();
StorageDescriptor sd = hiveTablePartition.getStorageDescriptor();
jobConf.set(INPUT_DIR, sd.getLocation());
InputFormat mapredInputFormat;
@@ -128,18 +122,30 @@ public class HiveMapredSplitReader implements SplitReader {
}
this.selectedFields = selectedFields;
- this.partitionKeys = partitionKeys;
- converters = Arrays.stream(selectedFields)
+ this.converters = Arrays.stream(selectedFields)
.mapToObj(i -> fieldTypes[i])
.map(DataFormatConverters::getConverterForDataType)
.toArray(DataFormatConverters.DataFormatConverter[]::new);
this.hiveShim = hiveShim;
+
+ // construct reuse row
+ this.row = new GenericRowData(selectedFields.length);
+ // set partition columns
+ if (!partitionKeys.isEmpty()) {
+ for (int i = 0; i < selectedFields.length; i++) {
+ if (selectedFields[i] >= structFields.size()) {
+ String partition = partitionKeys.get(selectedFields[i] - structFields.size());
+ row.setField(i, converters[i].toInternal(hiveTablePartition.getPartitionSpec().get(partition)));
+ }
+ }
+ }
}
@Override
public boolean reachedEnd() throws IOException {
if (!fetched) {
- fetchNext();
+ hasNext = recordReader.next(key, value);
+ fetched = true;
}
return !hasNext;
}
@@ -150,8 +156,6 @@ public class HiveMapredSplitReader implements SplitReader {
if (reachedEnd()) {
return null;
}
- final GenericRowData row = reuse instanceof GenericRowData ?
- (GenericRowData) reuse : new GenericRowData(selectedFields.length);
try {
//Use HiveDeserializer to deserialize an object out of a Writable blob
Object hiveRowStruct = deserializer.deserialize(value);
@@ -168,27 +172,10 @@ public class HiveMapredSplitReader implements SplitReader {
LOG.error("Error happens when converting hive data type to flink data type.");
throw new FlinkHiveException(e);
}
- if (!rowReused) {
- // set partition columns
- if (!partitionKeys.isEmpty()) {
- for (int i = 0; i < selectedFields.length; i++) {
- if (selectedFields[i] >= structFields.size()) {
- String partition = partitionKeys.get(selectedFields[i] - structFields.size());
- row.setField(i, converters[i].toInternal(hiveTablePartition.getPartitionSpec().get(partition)));
- }
- }
- }
- rowReused = true;
- }
this.fetched = false;
return row;
}
- private void fetchNext() throws IOException {
- hasNext = recordReader.next(key, value);
- fetched = true;
- }
-
@Override
public void close() throws IOException {
if (this.recordReader != null) {
diff --git a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormat.java b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormat.java
index eb1fb41..29e890b 100644
--- a/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormat.java
+++ b/flink-connectors/flink-connector-hive/src/main/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormat.java
@@ -18,6 +18,7 @@
package org.apache.flink.connectors.hive.read;
+import org.apache.flink.annotation.VisibleForTesting;
import org.apache.flink.api.common.functions.RuntimeContext;
import org.apache.flink.api.common.io.FileInputFormat;
import org.apache.flink.configuration.Configuration;
@@ -52,16 +53,26 @@ public class HiveTableFileInputFormat extends FileInputFormat<RowData> {
@Override
public void open(FileInputSplit fileSplit) throws IOException {
- URI uri = fileSplit.getPath().toUri();
HiveTableInputSplit split = new HiveTableInputSplit(
fileSplit.getSplitNumber(),
- new FileSplit(new Path(uri), fileSplit.getStart(), fileSplit.getLength(), (String[]) null),
+ toHadoopFileSplit(fileSplit),
inputFormat.getJobConf(),
- hiveTablePartition
- );
+ hiveTablePartition);
inputFormat.open(split);
}
+ @VisibleForTesting
+ static FileSplit toHadoopFileSplit(FileInputSplit fileSplit) throws IOException {
+ URI uri = fileSplit.getPath().toUri();
+ long length = fileSplit.getLength();
+ // Hadoop FileSplit should not have -1 length.
+ if (length == -1) {
+ length = fileSplit.getPath().getFileSystem().getFileStatus(fileSplit.getPath()).getLen() -
+ fileSplit.getStart();
+ }
+ return new FileSplit(new Path(uri), fileSplit.getStart(), length, (String[]) null);
+ }
+
@Override
public boolean reachedEnd() throws IOException {
return inputFormat.reachedEnd();
diff --git a/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/HiveTableSourceITCase.java b/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/HiveTableSourceITCase.java
index 32a5cd4..5170920 100644
--- a/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/HiveTableSourceITCase.java
+++ b/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/HiveTableSourceITCase.java
@@ -473,7 +473,7 @@ public class HiveTableSourceITCase extends BatchAbstractTestBase {
" b STRING" +
") PARTITIONED BY (ts STRING) TBLPROPERTIES (" +
"'streaming-source.enable'='true'," +
- "'streaming-source.monitor-interval'='100ms'," +
+ "'streaming-source.monitor-interval'='1s'," +
"'streaming-source.consume-order'='partition-time'" +
")");
@@ -502,6 +502,7 @@ public class HiveTableSourceITCase extends BatchAbstractTestBase {
}
HiveTestUtils.createTextTableInserter(hiveShell, dbName, tblName)
.addRow(new Object[]{i, String.valueOf(i)})
+ .addRow(new Object[]{i, i + "_copy"})
.commit("ts='2020-05-06 00:" + i + "0:00'");
}
};
@@ -514,10 +515,15 @@ public class HiveTableSourceITCase extends BatchAbstractTestBase {
List<String> expected = Arrays.asList(
"+I(0,0,2020-05-06 00:00:00)",
"+I(1,1,2020-05-06 00:10:00)",
+ "+I(1,1_copy,2020-05-06 00:10:00)",
"+I(2,2,2020-05-06 00:20:00)",
+ "+I(2,2_copy,2020-05-06 00:20:00)",
"+I(3,3,2020-05-06 00:30:00)",
+ "+I(3,3_copy,2020-05-06 00:30:00)",
"+I(4,4,2020-05-06 00:40:00)",
- "+I(5,5,2020-05-06 00:50:00)"
+ "+I(4,4_copy,2020-05-06 00:40:00)",
+ "+I(5,5,2020-05-06 00:50:00)",
+ "+I(5,5_copy,2020-05-06 00:50:00)"
);
List<String> results = sink.getJavaAppendResults();
results.sort(String::compareTo);
diff --git a/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscoveryTest.java b/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscoveryTest.java
new file mode 100644
index 0000000..ad514d4
--- /dev/null
+++ b/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/read/DirectoryMonitorDiscoveryTest.java
@@ -0,0 +1,90 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.connectors.hive.read;
+
+import org.apache.flink.api.java.tuple.Tuple2;
+import org.apache.flink.table.data.TimestampData;
+
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.hive.metastore.api.Partition;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.sql.Timestamp;
+import java.util.List;
+import java.util.Optional;
+import java.util.function.Supplier;
+
+/**
+ * Test for {@link DirectoryMonitorDiscovery}.
+ */
+public class DirectoryMonitorDiscoveryTest {
+
+ private static FileStatus status(String time) {
+ return new FileStatus(0L, false, 0, 0L, Timestamp.valueOf(time).getTime(), 0L, null, null, null, new Path("/tmp/dummy"));
+ }
+
+ @Test
+ public void testUTC() {
+ long previousTimestamp = TimestampData.fromTimestamp(Timestamp.valueOf("2020-05-06 12:22:00")).getMillisecond();
+ FileStatus[] statuses = new FileStatus[] {
+ status("2020-05-06 12:20:00"),
+ status("2020-05-06 12:21:00"),
+ status("2020-05-06 12:22:00"),
+ status("2020-05-06 12:23:00"),
+ status("2020-05-06 12:24:00")};
+ List<Tuple2<List<String>, Long>> parts = DirectoryMonitorDiscovery.suitablePartitions(
+ new PartitionDiscovery.Context() {
+
+ @Override
+ public List<String> partitionKeys() {
+ return null;
+ }
+
+ @Override
+ public Optional<Partition> getPartition(List<String> partValues) {
+ return Optional.empty();
+ }
+
+ @Override
+ public FileSystem fileSystem() {
+ return null;
+ }
+
+ @Override
+ public Path tableLocation() {
+ return null;
+ }
+
+ @Override
+ public long extractTimestamp(List<String> partKeys, List<String> partValues,
+ Supplier<Long> fileTime) {
+ return fileTime.get();
+ }
+ },
+ previousTimestamp,
+ statuses);
+ Assert.assertEquals(3, parts.size());
+ Assert.assertEquals("2020-05-06T12:22", TimestampData.fromEpochMillis(parts.get(0).f1).toString());
+ Assert.assertEquals("2020-05-06T12:23", TimestampData.fromEpochMillis(parts.get(1).f1).toString());
+ Assert.assertEquals("2020-05-06T12:24", TimestampData.fromEpochMillis(parts.get(2).f1).toString());
+ }
+}
diff --git a/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormatTest.java b/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormatTest.java
new file mode 100644
index 0000000..9a4fb37
--- /dev/null
+++ b/flink-connectors/flink-connector-hive/src/test/java/org/apache/flink/connectors/hive/read/HiveTableFileInputFormatTest.java
@@ -0,0 +1,50 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.connectors.hive.read;
+
+import org.apache.flink.core.fs.FileInputSplit;
+import org.apache.flink.core.fs.Path;
+import org.apache.flink.util.FileUtils;
+
+import org.apache.hadoop.mapred.FileSplit;
+import org.junit.Assert;
+import org.junit.ClassRule;
+import org.junit.Test;
+import org.junit.rules.TemporaryFolder;
+
+import java.io.File;
+import java.io.IOException;
+
+/**
+ * Test for {@link HiveTableFileInputFormat}.
+ */
+public class HiveTableFileInputFormatTest {
+
+ @ClassRule
+ public static final TemporaryFolder TEMPORARY_FOLDER = new TemporaryFolder();
+
+ @Test
+ public void testSplit() throws IOException {
+ File file = TEMPORARY_FOLDER.newFile();
+ FileUtils.writeFileUtf8(file, "hahahahahahaha");
+ FileInputSplit split = new FileInputSplit(0, new Path(file.getPath()), 0, -1, null);
+ FileSplit fileSplit = HiveTableFileInputFormat.toHadoopFileSplit(split);
+ Assert.assertEquals(14, fileSplit.getLength());
+ }
+}