test/org/apache/pig/test/TestNumberOfReducers.java - pig - Git at Google

 /*
  * 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.pig.test;

 import java.io.File;
 import java.io.FileWriter;
 import java.io.IOException;
 import java.io.PrintWriter;

 import org.apache.pig.PigRunner;
 import org.apache.pig.PigServer;
 import org.apache.pig.impl.PigContext;
 import org.apache.pig.tools.pigstats.PigStats;
 import org.apache.pig.tools.pigstats.mapreduce.MRJobStats;
 import org.junit.AfterClass;
 import org.junit.BeforeClass;
 import org.junit.Test;
 import org.junit.runner.RunWith;
 import org.junit.runners.JUnit4;

 import static org.junit.Assert.assertEquals;
 import static org.junit.Assert.assertNull;
 import static org.junit.Assert.assertTrue;

 /**
  * This class tests whether the number of reducers is correctly set for queries
  * involving a sample job. Note the sample job should use the actual #reducers
  * of the following order-by or skew-join job to generate the partition file,
  * and the challenge is that whether we can get the actual #reducer of the
  * order-by/skew-join job in advance.
  */
 @RunWith(JUnit4.class)
 public class TestNumberOfReducers {

     private static final String LOCAL_INPUT_FILE = "test/org/apache/pig/test/data/passwd";
     private static final String HDFS_INPUT_FILE = "passwd";
     private static final String OUTPUT_FILE = "output";
     private static final String PIG_FILE = "test.pig";

     static PigContext pc;
     static PigServer pigServer;
     private static MiniGenericCluster cluster;

     @BeforeClass
     public static void setUpBeforeClass() throws Exception {
         cluster = MiniGenericCluster.buildCluster();

         Util.copyFromLocalToCluster(cluster, LOCAL_INPUT_FILE, HDFS_INPUT_FILE);
     }

     @AfterClass
     public static void oneTimeTearDown() throws Exception {
         cluster.shutDown();
     }

     /**
      * Specifying bytes_per_reducer, default_parallel and parallel, verifying
      * whether it uses the actual #reducers as actual_parallel.
      *
      * Note the input file size is 555B.
      *
      */
     private void verifyOrderBy(int bytes_per_reducer, int default_parallel,
             int parallel, int actual_parallel) throws IOException {
         PrintWriter w = new PrintWriter(new FileWriter(PIG_FILE));

         // the file is 555B
         w.println("SET pig.exec.reducers.bytes.per.reducer " + bytes_per_reducer + ";");

         w.println("SET default_parallel " + default_parallel + ";");
         w.println("A = load '" + HDFS_INPUT_FILE + "';");
         w.print("B = order A by $0 ");
         if (parallel > 0)
             w.print("parallel " + parallel);
         w.println(";");
         w.println("store B into '" + OUTPUT_FILE + "';");
         w.close();

         doTest(bytes_per_reducer, default_parallel, parallel, actual_parallel);
     }

     /**
      * Specifying bytes_per_reducer, default_parallel and parallel, verifying
      * whether it uses the actual #reducers as actual_parallel.
      *
      * Note the input file size is 555B.
      *
      */
     private void verifySkewJoin(int bytes_per_reducer, int default_parallel,
             int parallel, int actual_parallel) throws IOException {
         PrintWriter w = new PrintWriter(new FileWriter(PIG_FILE));

         // the file is 555B
         w.println("SET pig.exec.reducers.bytes.per.reducer " + bytes_per_reducer + ";");
         w.println("SET default_parallel " + default_parallel + ";");
         w.println("A1 = load '" + HDFS_INPUT_FILE + "';");
         w.println("A2 = load '" + HDFS_INPUT_FILE + "';");
         w.print("B = join A1 by $0,  A2 by $0 using 'skewed' ");
         if (parallel > 0)
             w.print("parallel " + parallel);
         w.println(";");
         w.println("store B into '" + OUTPUT_FILE + "';");
         w.close();

         doTest(bytes_per_reducer, default_parallel, parallel, actual_parallel);
     }

     private void doTest(double bytes_per_reducer, int default_parallel,
                         int parallel, int actual_parallel) throws IOException {
         try {
             String[] args = { PIG_FILE };
             PigStats stats = PigRunner.run(args, null);

             assertTrue(stats.isSuccessful());

             // get the skew-join job stat
             MRJobStats js = (MRJobStats) stats.getJobGraph().getSinks().get(0);
             assertEquals(actual_parallel, js.getNumberReduces());

             // estimation should only kick in if parallel and default_parallel are not set
             long estimatedReducers = -1;
             if (parallel < 1 && default_parallel < 1) {
                 double fileSize = (double)(new File("test/org/apache/pig/test/data/passwd").length());
                 int inputFiles = js.getInputs().size();
                 estimatedReducers = Math.min((long)Math.ceil(fileSize/(double)bytes_per_reducer) * inputFiles, 999);
             }

             Util.assertParallelValues(default_parallel, parallel, estimatedReducers,
                     actual_parallel, js.getInputs().get(0).getConf());

         } catch (Exception e) {
             assertNull("Exception thrown during verifySkewJoin", e);
         } finally {
             new File(PIG_FILE).delete();
             Util.deleteFile(cluster, OUTPUT_FILE);
         }
     }

     /**
      * For 550B file and 300B per reducer, it estimates 2 #reducers.
      * @throws Exception
      */
     @Test
     public void testOrderByEstimate() throws Exception{
         verifyOrderBy(300, -1, -1, 2);
     }

     /**
      * For two 550B files and 600B per reducer, it estimates 2 #reducers.
      * @throws Exception
      */
     @Test
     public void testSkewJoinEstimate() throws Exception{
         verifySkewJoin(600, -1, -1, 2);
     }


     /**
      * Pig will estimate 2 reducers but we specify parallel 1, so it should use
      * 1 reducer.
      *
      * @throws Exception
      */
     @Test
     public void testOrderByEstimate2Parallel1() throws Exception {
         verifyOrderBy(300, -1, 1, 1);
     }

     /**
      * Pig will estimate 2 reducers but we specify parallel 1, so it should use
      * 1 reducer.
      *
      * @throws Exception
      */
     @Test
     public void testSkewJoinEstimate2Parallel1() throws Exception {
         verifySkewJoin(600, -1, 1, 1);
     }

     /**
      * Pig will estimate 2 reducers but we specify default parallel 1, so it
      * should use 1 reducer.
      *
      * @throws Exception
      */
     @Test
     public void testOrderByEstimate2Default1() throws Exception {
         verifyOrderBy(300,  1, -1, 1);
     }

     /**
      * Pig will estimate 2 reducers but we specify default parallel 1, so it
      * should use 1 reducer.
      *
      * @throws Exception
      */
     @Test
     public void testSkewJoinEstimate2Default1() throws Exception {
         verifySkewJoin(600, 1, -1, 1);
     }

     /**
      * Pig will estimate 2 reducers but we specify parallel 4, so it should use
      * 4 reducer.
      *
      * TODO we need verify that the sample job generates 4 partitions instead of
      * 2.
      *
      * @throws Exception
      */
     @Test
     public void testOrderByEstimate2Parallel4() throws Exception {
         verifyOrderBy(300, -1, 4, 4);
     }

     /**
      * Pig will estimate 2 reducers but we specify parallel 4, so it should use
      * 4 reducer.
      *
      * TODO we need verify that the sample job generates 4 partitions instead of
      * 2.
      *
      * @throws Exception
      */
     @Test
     public void testSkewJoinEstimate2Parallel4() throws Exception {
         verifySkewJoin(600, -1, 4, 4);
     }

     /**
      * Pig will estimate 6 reducers but we specify default parallel 2, so it
      * should use 2 reducer. Also, the sampler should generate 2 partitions,
      * instead of 6, otherwise the next job will fail.
      *
      * @throws Exception
      */
     @Test
     public void testOrderByEstimate6Default2() throws Exception {
         verifyOrderBy(100, 2, -1, 2);
     }

     /**
      * Pig will estimate 6 reducers but we specify default parallel 2, so it
      * should use 2 reducer. Also, the sampler should generate 2 partitions,
      * instead of 6, otherwise the next job will fail.
      *
      * @throws Exception
      */
     @Test
     public void testSkewJoinEstimate6Default2() throws Exception {
         verifySkewJoin(200, 2, -1, 2);
     }
 }
	/*
	* 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.pig.test;

	import java.io.File;
	import java.io.FileWriter;
	import java.io.IOException;
	import java.io.PrintWriter;

	import org.apache.pig.PigRunner;
	import org.apache.pig.PigServer;
	import org.apache.pig.impl.PigContext;
	import org.apache.pig.tools.pigstats.PigStats;
	import org.apache.pig.tools.pigstats.mapreduce.MRJobStats;
	import org.junit.AfterClass;
	import org.junit.BeforeClass;
	import org.junit.Test;
	import org.junit.runner.RunWith;
	import org.junit.runners.JUnit4;

	import static org.junit.Assert.assertEquals;
	import static org.junit.Assert.assertNull;
	import static org.junit.Assert.assertTrue;

	/**
	* This class tests whether the number of reducers is correctly set for queries
	* involving a sample job. Note the sample job should use the actual #reducers
	* of the following order-by or skew-join job to generate the partition file,
	* and the challenge is that whether we can get the actual #reducer of the
	* order-by/skew-join job in advance.
	*/
	@RunWith(JUnit4.class)
	public class TestNumberOfReducers {

	private static final String LOCAL_INPUT_FILE = "test/org/apache/pig/test/data/passwd";
	private static final String HDFS_INPUT_FILE = "passwd";
	private static final String OUTPUT_FILE = "output";
	private static final String PIG_FILE = "test.pig";

	static PigContext pc;
	static PigServer pigServer;
	private static MiniGenericCluster cluster;

	@BeforeClass
	public static void setUpBeforeClass() throws Exception {
	cluster = MiniGenericCluster.buildCluster();

	Util.copyFromLocalToCluster(cluster, LOCAL_INPUT_FILE, HDFS_INPUT_FILE);
	}

	@AfterClass
	public static void oneTimeTearDown() throws Exception {
	cluster.shutDown();
	}

	/**
	* Specifying bytes_per_reducer, default_parallel and parallel, verifying
	* whether it uses the actual #reducers as actual_parallel.
	*
	* Note the input file size is 555B.
	*
	*/
	private void verifyOrderBy(int bytes_per_reducer, int default_parallel,
	int parallel, int actual_parallel) throws IOException {
	PrintWriter w = new PrintWriter(new FileWriter(PIG_FILE));

	// the file is 555B
	w.println("SET pig.exec.reducers.bytes.per.reducer " + bytes_per_reducer + ";");

	w.println("SET default_parallel " + default_parallel + ";");
	w.println("A = load '" + HDFS_INPUT_FILE + "';");
	w.print("B = order A by $0 ");
	if (parallel > 0)
	w.print("parallel " + parallel);
	w.println(";");
	w.println("store B into '" + OUTPUT_FILE + "';");
	w.close();

	doTest(bytes_per_reducer, default_parallel, parallel, actual_parallel);
	}

	/**
	* Specifying bytes_per_reducer, default_parallel and parallel, verifying
	* whether it uses the actual #reducers as actual_parallel.
	*
	* Note the input file size is 555B.
	*
	*/
	private void verifySkewJoin(int bytes_per_reducer, int default_parallel,
	int parallel, int actual_parallel) throws IOException {
	PrintWriter w = new PrintWriter(new FileWriter(PIG_FILE));

	// the file is 555B
	w.println("SET pig.exec.reducers.bytes.per.reducer " + bytes_per_reducer + ";");
	w.println("SET default_parallel " + default_parallel + ";");
	w.println("A1 = load '" + HDFS_INPUT_FILE + "';");
	w.println("A2 = load '" + HDFS_INPUT_FILE + "';");
	w.print("B = join A1 by $0, A2 by $0 using 'skewed' ");
	if (parallel > 0)
	w.print("parallel " + parallel);
	w.println(";");
	w.println("store B into '" + OUTPUT_FILE + "';");
	w.close();

	doTest(bytes_per_reducer, default_parallel, parallel, actual_parallel);
	}

	private void doTest(double bytes_per_reducer, int default_parallel,
	int parallel, int actual_parallel) throws IOException {
	try {
	String[] args = { PIG_FILE };
	PigStats stats = PigRunner.run(args, null);

	assertTrue(stats.isSuccessful());

	// get the skew-join job stat
	MRJobStats js = (MRJobStats) stats.getJobGraph().getSinks().get(0);
	assertEquals(actual_parallel, js.getNumberReduces());

	// estimation should only kick in if parallel and default_parallel are not set
	long estimatedReducers = -1;
	if (parallel < 1 && default_parallel < 1) {
	double fileSize = (double)(new File("test/org/apache/pig/test/data/passwd").length());
	int inputFiles = js.getInputs().size();
	estimatedReducers = Math.min((long)Math.ceil(fileSize/(double)bytes_per_reducer) * inputFiles, 999);
	}

	Util.assertParallelValues(default_parallel, parallel, estimatedReducers,
	actual_parallel, js.getInputs().get(0).getConf());

	} catch (Exception e) {
	assertNull("Exception thrown during verifySkewJoin", e);
	} finally {
	new File(PIG_FILE).delete();
	Util.deleteFile(cluster, OUTPUT_FILE);
	}
	}

	/**
	* For 550B file and 300B per reducer, it estimates 2 #reducers.
	* @throws Exception
	*/
	@Test
	public void testOrderByEstimate() throws Exception{
	verifyOrderBy(300, -1, -1, 2);
	}

	/**
	* For two 550B files and 600B per reducer, it estimates 2 #reducers.
	* @throws Exception
	*/
	@Test
	public void testSkewJoinEstimate() throws Exception{
	verifySkewJoin(600, -1, -1, 2);
	}


	/**
	* Pig will estimate 2 reducers but we specify parallel 1, so it should use
	* 1 reducer.
	*
	* @throws Exception
	*/
	@Test
	public void testOrderByEstimate2Parallel1() throws Exception {
	verifyOrderBy(300, -1, 1, 1);
	}

	/**
	* Pig will estimate 2 reducers but we specify parallel 1, so it should use
	* 1 reducer.
	*
	* @throws Exception
	*/
	@Test
	public void testSkewJoinEstimate2Parallel1() throws Exception {
	verifySkewJoin(600, -1, 1, 1);
	}

	/**
	* Pig will estimate 2 reducers but we specify default parallel 1, so it
	* should use 1 reducer.
	*
	* @throws Exception
	*/
	@Test
	public void testOrderByEstimate2Default1() throws Exception {
	verifyOrderBy(300, 1, -1, 1);
	}

	/**
	* Pig will estimate 2 reducers but we specify default parallel 1, so it
	* should use 1 reducer.
	*
	* @throws Exception
	*/
	@Test
	public void testSkewJoinEstimate2Default1() throws Exception {
	verifySkewJoin(600, 1, -1, 1);
	}

	/**
	* Pig will estimate 2 reducers but we specify parallel 4, so it should use
	* 4 reducer.
	*
	* TODO we need verify that the sample job generates 4 partitions instead of
	* 2.
	*
	* @throws Exception
	*/
	@Test
	public void testOrderByEstimate2Parallel4() throws Exception {
	verifyOrderBy(300, -1, 4, 4);
	}

	/**
	* Pig will estimate 2 reducers but we specify parallel 4, so it should use
	* 4 reducer.
	*
	* TODO we need verify that the sample job generates 4 partitions instead of
	* 2.
	*
	* @throws Exception
	*/
	@Test
	public void testSkewJoinEstimate2Parallel4() throws Exception {
	verifySkewJoin(600, -1, 4, 4);
	}

	/**
	* Pig will estimate 6 reducers but we specify default parallel 2, so it
	* should use 2 reducer. Also, the sampler should generate 2 partitions,
	* instead of 6, otherwise the next job will fail.
	*
	* @throws Exception
	*/
	@Test
	public void testOrderByEstimate6Default2() throws Exception {
	verifyOrderBy(100, 2, -1, 2);
	}

	/**
	* Pig will estimate 6 reducers but we specify default parallel 2, so it
	* should use 2 reducer. Also, the sampler should generate 2 partitions,
	* instead of 6, otherwise the next job will fail.
	*
	* @throws Exception
	*/
	@Test
	public void testSkewJoinEstimate6Default2() throws Exception {
	verifySkewJoin(200, 2, -1, 2);
	}
	}