You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@parquet.apache.org by GitBox <gi...@apache.org> on 2020/05/21 21:11:47 UTC
[GitHub] [parquet-mr] shangxinli opened a new pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
shangxinli opened a new pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793
Make sure you have checked _all_ steps below.
### Jira
- [ ] My PR addresses the following [Parquet Jira](https://issues.apache.org/jira/browse/PARQUET/) issues and references them in the PR title. For example, "PARQUET-1234: My Parquet PR"
- https://issues.apache.org/jira/browse/PARQUET-XXX
- In case you are adding a dependency, check if the license complies with the [ASF 3rd Party License Policy](https://www.apache.org/legal/resolved.html#category-x).
### Tests
- [ ] My PR adds the following unit tests __OR__ does not need testing for this extremely good reason:
### Commits
- [ ] My commits all reference Jira issues in their subject lines. In addition, my commits follow the guidelines from "[How to write a good git commit message](http://chris.beams.io/posts/git-commit/)":
1. Subject is separated from body by a blank line
1. Subject is limited to 50 characters (not including Jira issue reference)
1. Subject does not end with a period
1. Subject uses the imperative mood ("add", not "adding")
1. Body wraps at 72 characters
1. Body explains "what" and "why", not "how"
### Documentation
- [ ] In case of new functionality, my PR adds documentation that describes how to use it.
- All the public functions and the classes in the PR contain Javadoc that explain what it does
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] dbtsai commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
dbtsai commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-637193519
+1 @shangxinli and thank you for this contribution.
This will allow users who are on order versions of hadoop that don't support native ZSTD to use ZSTD compression in Parquet, and also, users don't have to go through the very complicated hadoop native installation. For developers, we will be able to easily test this out in different local envs.
cc @rdblue
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] luben commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
luben commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-639394974
@shangxinli : I haven't benchmarked
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432667013
##########
File path: parquet-hadoop/README.md
##########
@@ -324,9 +324,20 @@ ParquetInputFormat to materialize records. It should be a the descendant class o
**Property:** `parquet.read.schema`
**Description:** The read projection schema.
-
## Class: UnmaterializableRecordCounter
**Property:** `parquet.read.bad.record.threshold`
**Description:** The percentage of bad records to tolerate.
**Default value:** `0`
+
+## Class: ZstandardCodec
+
+**Property:** `parquet.compression.codec.zstd.level`
+**Description:** The compression level of ZSTD. The valid range is 1~22. Generally the higher compression level, the higher compression ratio can be achieved, but the writing time will be longer.
Review comment:
I see. I added double space in the end of Property, Description sentences. I checked it in Intellij and it works. I also see the above sections used the trailing double space.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432709258
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,146 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 18);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 4);
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf);
+ assert(mapRedJob.isSuccessful());
+ }
+
+ private RunningJob runMapReduceJob(CompressionCodecName codec, JobConf jobConf, Configuration conf) throws IOException, ClassNotFoundException, InterruptedException {
Review comment:
I like the idea to compare the two files with different compression levels. Only thing is that it is not 100% guarantee the higher level always gets higher compression ratio although the chance of that is very low. In our test data, it seems it is OK though. So I will just go ahead to add the file size comparison.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r429560767
##########
File path: parquet-hadoop/src/main/java/org/apache/parquet/hadoop/codec/ZstdCodec.java
##########
@@ -0,0 +1,112 @@
+/*
+ * 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.parquet.hadoop.codec;
+
+import org.apache.hadoop.conf.Configurable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.io.compress.CompressionCodec;
+import org.apache.hadoop.io.compress.CompressionInputStream;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.io.compress.Compressor;
+import org.apache.hadoop.io.compress.Decompressor;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
+
+/**
+ * ZSTD compression codec for Parquet. We do not use the default hadoop
+ * one because it requires 1) to set up hadoop on local develop machine;
+ * 2) to upgrade hadoop to the newer version to have ZSTD support which is
+ * more cumbersome than upgrading parquet version.
+ *
+ * This implementation relies on ZSTD JNI(https://github.com/luben/zstd-jni)
+ * which is already a dependency for Parquet. ZSTD JNI ZstdOutputStream and
+ * ZstdInputStream use Zstd internally. So no need to create compressor and
+ * decompressor in ZstdCodec.
+ */
+public class ZstdCodec implements Configurable, CompressionCodec {
+
+ public final static String PARQUET_COMPRESS_ZSTD_LEVEL = "parquet.compression.codec.zstd.level";
+ public final static int DEFAULT_PARQUET_COMPRESS_ZSTD_LEVEL = 3;
+ public final static String PARQUET_COMPRESS_ZSTD_WORKERS = "parquet.compression.codec.zstd.workers";
+ public final static int DEFAULTPARQUET_COMPRESS_ZSTD_WORKERS = 0;
Review comment:
Sure
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r434002793
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,164 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
Review comment:
Added. Thanks.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] luben commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
luben commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-634146458
LGTM
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] gszadovszky commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
gszadovszky commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r433675252
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,146 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 18);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 4);
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf);
+ assert(mapRedJob.isSuccessful());
+ }
+
+ private RunningJob runMapReduceJob(CompressionCodecName codec, JobConf jobConf, Configuration conf) throws IOException, ClassNotFoundException, InterruptedException {
Review comment:
Yes, I know, there is no 100% guarantee in normal circumstances. But if you write the test in a deterministic way (as it should be by e.g. using a fix seed for random generation) then without changing the data and the compression codec it shall be fine. I worth a comment, though about it is not the nicest way of testing the work of the properties but currently that's seems to be the only one.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] gszadovszky commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
gszadovszky commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432376409
##########
File path: parquet-hadoop/README.md
##########
@@ -324,9 +324,20 @@ ParquetInputFormat to materialize records. It should be a the descendant class o
**Property:** `parquet.read.schema`
**Description:** The read projection schema.
-
## Class: UnmaterializableRecordCounter
**Property:** `parquet.read.bad.record.threshold`
**Description:** The percentage of bad records to tolerate.
**Default value:** `0`
+
+## Class: ZstandardCodec
+
+**Property:** `parquet.compression.codec.zstd.level`
+**Description:** The compression level of ZSTD. The valid range is 1~22. Generally the higher compression level, the higher compression ratio can be achieved, but the writing time will be longer.
Review comment:
I was trying to say that _Property_, _Description_ and _Default value_ should be separate paragraphs. Currently they are not, the _Description_ is rendered just after _Property_ in the same line even if in the markdown they are separated. You should use one of the techniques described under the link to force the new line there.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-634988634
@gszadovszky Do you have time for another look?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-639245331
> @shangxinli do we have benchmark comparing to native hadoop codec both in size and speed? Thanks.
Hi @dbtsai, I didn't because I don't have Hadoop host installed with ZSTD. @luben, did you ever compare it with Hadoop ZSTD?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli edited a comment on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli edited a comment on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-634104930
@luben, Do you have time to review the code?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] gszadovszky commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
gszadovszky commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r429112028
##########
File path: parquet-hadoop/src/main/java/org/apache/parquet/hadoop/codec/ZstdCodec.java
##########
@@ -0,0 +1,112 @@
+/*
+ * 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.parquet.hadoop.codec;
+
+import org.apache.hadoop.conf.Configurable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.io.compress.CompressionCodec;
+import org.apache.hadoop.io.compress.CompressionInputStream;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.io.compress.Compressor;
+import org.apache.hadoop.io.compress.Decompressor;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
+
+/**
+ * ZSTD compression codec for Parquet. We do not use the default hadoop
+ * one because it requires 1) to set up hadoop on local develop machine;
Review comment:
```suggestion
* one because it requires 1) to set up hadoop on local development machine;
```
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstdCodec.java
##########
@@ -0,0 +1,74 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.hadoop.codec.ZstdCodec;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.Random;
+
+public class TestZstdCodec {
Review comment:
I tried to find the code part where we set the hadoop conf to the codec but could not find it. Please, write a high level test where you set compression level and workers in the hadoop conf and executes a file write via e.g. an MR job.
##########
File path: parquet-hadoop/src/main/java/org/apache/parquet/hadoop/codec/ZstdCodec.java
##########
@@ -0,0 +1,112 @@
+/*
+ * 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.parquet.hadoop.codec;
+
+import org.apache.hadoop.conf.Configurable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.io.compress.CompressionCodec;
+import org.apache.hadoop.io.compress.CompressionInputStream;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.io.compress.Compressor;
+import org.apache.hadoop.io.compress.Decompressor;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
+
+/**
+ * ZSTD compression codec for Parquet. We do not use the default hadoop
+ * one because it requires 1) to set up hadoop on local develop machine;
+ * 2) to upgrade hadoop to the newer version to have ZSTD support which is
+ * more cumbersome than upgrading parquet version.
+ *
+ * This implementation relies on ZSTD JNI(https://github.com/luben/zstd-jni)
+ * which is already a dependency for Parquet. ZSTD JNI ZstdOutputStream and
+ * ZstdInputStream use Zstd internally. So no need to create compressor and
+ * decompressor in ZstdCodec.
+ */
+public class ZstdCodec implements Configurable, CompressionCodec {
+
+ public final static String PARQUET_COMPRESS_ZSTD_LEVEL = "parquet.compression.codec.zstd.level";
+ public final static int DEFAULT_PARQUET_COMPRESS_ZSTD_LEVEL = 3;
+ public final static String PARQUET_COMPRESS_ZSTD_WORKERS = "parquet.compression.codec.zstd.workers";
+ public final static int DEFAULTPARQUET_COMPRESS_ZSTD_WORKERS = 0;
Review comment:
Please, also update the documentation in the [README](https://github.com/apache/parquet-mr/blob/master/parquet-hadoop/README.md).
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-634104930
@karavelov, Do you have time to review the code?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] dongjoon-hyun commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
dongjoon-hyun commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-636231687
Thank you, @shangxinli and all!
cc @dbtsai
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r434001031
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,164 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ long fileSizeLowLevel = runMrWithConf(1);
+ // Clear the cache so that a new codec can be created with new configuration
+ CodecFactory.CODEC_BY_NAME.clear();
+ long fileSizeHighLevel = runMrWithConf(22);
+ assert (fileSizeLowLevel > fileSizeHighLevel);
Review comment:
Sounds good!
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,164 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ long fileSizeLowLevel = runMrWithConf(1);
+ // Clear the cache so that a new codec can be created with new configuration
+ CodecFactory.CODEC_BY_NAME.clear();
+ long fileSizeHighLevel = runMrWithConf(22);
+ assert (fileSizeLowLevel > fileSizeHighLevel);
+ }
+
+ private long runMrWithConf(int level) throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, level);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_WORKERS, 4);
+ Path path = new Path(Files.createTempDirectory("zstd" + level).toAbsolutePath().toString());
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf, path);
+ assert(mapRedJob.isSuccessful());
Review comment:
fixed
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432142474
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,146 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 18);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 4);
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf);
+ assert(mapRedJob.isSuccessful());
+ }
+
+ private RunningJob runMapReduceJob(CompressionCodecName codec, JobConf jobConf, Configuration conf) throws IOException, ClassNotFoundException, InterruptedException {
Review comment:
I set breakpoint and checked it. But I don't know how to get back the level/workers at the MR job. Any ideal how to do that? I also verified it in the standalone Spark application. Do you think that is sufficient?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] gszadovszky merged pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
gszadovszky merged pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r434007162
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,146 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 18);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 4);
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf);
+ assert(mapRedJob.isSuccessful());
+ }
+
+ private RunningJob runMapReduceJob(CompressionCodecName codec, JobConf jobConf, Configuration conf) throws IOException, ClassNotFoundException, InterruptedException {
Review comment:
Thanks.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432118673
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,146 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 18);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 4);
Review comment:
Good catch
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r429560783
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstdCodec.java
##########
@@ -0,0 +1,74 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.hadoop.codec.ZstdCodec;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.Random;
+
+public class TestZstdCodec {
Review comment:
Added test for conf setting with MR
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] gszadovszky commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
gszadovszky commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r433723783
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,164 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ long fileSizeLowLevel = runMrWithConf(1);
+ // Clear the cache so that a new codec can be created with new configuration
+ CodecFactory.CODEC_BY_NAME.clear();
+ long fileSizeHighLevel = runMrWithConf(22);
+ assert (fileSizeLowLevel > fileSizeHighLevel);
Review comment:
Please use the JUnit framework `assert` functions instead of the `assert` keyword. The [`assert` keyword](https://docs.oracle.com/javase/8/docs/technotes/guides/language/assert.html) is not for unit testing and it is not 100% guaranteed that the assertions are enabled during testing.
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,164 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
Review comment:
Please, add some comments that the intent of this test is to verify that the properties are passed through from the config to the codec.
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,164 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ long fileSizeLowLevel = runMrWithConf(1);
+ // Clear the cache so that a new codec can be created with new configuration
+ CodecFactory.CODEC_BY_NAME.clear();
+ long fileSizeHighLevel = runMrWithConf(22);
+ assert (fileSizeLowLevel > fileSizeHighLevel);
+ }
+
+ private long runMrWithConf(int level) throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, level);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_WORKERS, 4);
+ Path path = new Path(Files.createTempDirectory("zstd" + level).toAbsolutePath().toString());
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf, path);
+ assert(mapRedJob.isSuccessful());
Review comment:
Same as above
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] gszadovszky commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
gszadovszky commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432381728
##########
File path: parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestZstandardCodec.java
##########
@@ -0,0 +1,146 @@
+/*
+ * 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.parquet.hadoop;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.io.LongWritable;
+import org.apache.hadoop.io.Text;
+import org.apache.hadoop.io.compress.CompressionOutputStream;
+import org.apache.hadoop.mapred.JobClient;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.OutputCollector;
+import org.apache.hadoop.mapred.Reporter;
+import org.apache.hadoop.mapred.RunningJob;
+import org.apache.hadoop.mapred.TextInputFormat;
+import org.apache.parquet.bytes.BytesInput;
+import org.apache.parquet.example.data.Group;
+import org.apache.parquet.example.data.simple.SimpleGroupFactory;
+import org.apache.parquet.hadoop.codec.ZstandardCodec;
+import org.apache.parquet.hadoop.example.GroupWriteSupport;
+import org.apache.parquet.hadoop.mapred.DeprecatedParquetOutputFormat;
+import org.apache.parquet.hadoop.metadata.CompressionCodecName;
+import org.apache.parquet.schema.MessageTypeParser;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.nio.file.Files;
+import java.util.Random;
+
+public class TestZstandardCodec {
+
+ private final Path inputPath = new Path("src/test/java/org/apache/parquet/hadoop/example/TestInputOutputFormat.java");
+
+ @Test
+ public void testZstdCodec() throws IOException {
+ ZstandardCodec codec = new ZstandardCodec();
+ Configuration conf = new Configuration();
+ int[] levels = {1, 4, 7, 10, 13, 16, 19, 22};
+ int[] dataSizes = {0, 1, 10, 1024, 1024 * 1024};
+
+ for (int i = 0; i < levels.length; i++) {
+ conf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, levels[i]);
+ codec.setConf(conf);
+ for (int j = 0; j < dataSizes.length; j++) {
+ testZstd(codec, dataSizes[j]);
+ }
+ }
+ }
+
+ private void testZstd(ZstandardCodec codec, int dataSize) throws IOException {
+ byte[] data = new byte[dataSize];
+ (new Random()).nextBytes(data);
+ BytesInput compressedData = compress(codec, BytesInput.from(data));
+ BytesInput decompressedData = decompress(codec, compressedData, data.length);
+ Assert.assertArrayEquals(data, decompressedData.toByteArray());
+ }
+
+ private BytesInput compress(ZstandardCodec codec, BytesInput bytes) throws IOException {
+ ByteArrayOutputStream compressedOutBuffer = new ByteArrayOutputStream((int)bytes.size());
+ CompressionOutputStream cos = codec.createOutputStream(compressedOutBuffer, null);
+ bytes.writeAllTo(cos);
+ cos.close();
+ return BytesInput.from(compressedOutBuffer);
+ }
+
+ private BytesInput decompress(ZstandardCodec codec, BytesInput bytes, int uncompressedSize) throws IOException {
+ BytesInput decompressed;
+ InputStream is = codec.createInputStream(bytes.toInputStream(), null);
+ decompressed = BytesInput.from(is, uncompressedSize);
+ is.close();
+ return decompressed;
+ }
+
+ @Test
+ public void testZstdConfWithMr() throws Exception {
+ JobConf jobConf = new JobConf();
+ Configuration conf = new Configuration();
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 18);
+ jobConf.setInt(ZstandardCodec.PARQUET_COMPRESS_ZSTD_LEVEL, 4);
+ RunningJob mapRedJob = runMapReduceJob(CompressionCodecName.ZSTD, jobConf, conf);
+ assert(mapRedJob.isSuccessful());
+ }
+
+ private RunningJob runMapReduceJob(CompressionCodecName codec, JobConf jobConf, Configuration conf) throws IOException, ClassNotFoundException, InterruptedException {
Review comment:
Maybe, if we create two parquet files with different levels (e.g. 1 and 22) but with the exact same data we can expect that the larger level will generate smaller file?
I just want to ensure that the properties really arrive to the codec. If you think the related test requires too much effort or even not feasible I am fine if you can point me to the code path where the properties are passed through to the codecs.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] dbtsai commented on pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
dbtsai commented on pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#issuecomment-637195534
@shangxinli do we have benchmark comparing to native hadoop codec both in size and speed? Thanks.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [parquet-mr] shangxinli commented on a change in pull request #793: PARQUET-1866: Replace Hadoop ZSTD with JNI-ZSTD
Posted by GitBox <gi...@apache.org>.
shangxinli commented on a change in pull request #793:
URL: https://github.com/apache/parquet-mr/pull/793#discussion_r432118486
##########
File path: parquet-hadoop/README.md
##########
@@ -324,9 +324,20 @@ ParquetInputFormat to materialize records. It should be a the descendant class o
**Property:** `parquet.read.schema`
**Description:** The read projection schema.
-
## Class: UnmaterializableRecordCounter
**Property:** `parquet.read.bad.record.threshold`
**Description:** The percentage of bad records to tolerate.
**Default value:** `0`
+
+## Class: ZstandardCodec
+
+**Property:** `parquet.compression.codec.zstd.level`
+**Description:** The compression level of ZSTD. The valid range is 1~22. Generally the higher compression level, the higher compression ratio can be achieved, but the writing time will be longer.
Review comment:
Thanks for pointing out! I think it is OK to keep them in same line unless you have strong opinion we should have a separate line. I looked at the existing lines above and I see some of them are longer than mines.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org