Package org.apache.pig.data

Source Code of org.apache.pig.data.DefaultAbstractBag$EndBag

/*
* 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.data;

import java.io.BufferedOutputStream;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.DataOutputStream;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.util.Collection;
import java.util.Iterator;

import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.pig.PigException;
import org.apache.pig.PigWarning;
import org.apache.pig.backend.executionengine.ExecException;
import org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigHadoopLogger;
import org.apache.pig.backend.hadoop.executionengine.physicalLayer.PhysicalOperator;
import org.apache.pig.backend.hadoop.executionengine.physicalLayer.PigLogger;
import org.apache.pig.impl.util.BagFormat;
import org.apache.pig.impl.util.SpillableMemoryManager;
import org.apache.pig.tools.pigstats.PigStatusReporter;

/**
* Default implementation of DataBag.  This is the an abstract class used as a
* parent for all three of the types of data bags.
*/
@SuppressWarnings("serial")
public abstract class DefaultAbstractBag implements DataBag {

    private static final Log log = LogFactory.getLog(DataBag.class);

    // If we grow past 100K, may be worthwhile to register.
    private static final int SPILL_REGISTER_THRESHOLD = 100 * 1024;

    private static PigLogger pigLogger;

    private static InterSedes sedes = InterSedesFactory.getInterSedesInstance();
    // Container that holds the tuples. Actual object instantiated by
    // subclasses.
    protected Collection<Tuple> mContents;

    // Spill files we've created.  These need to be removed in finalize.
    protected FileList mSpillFiles;

    // Total size, including tuples on disk.  Stored here so we don't have
    // to run through the disk when people ask.
    protected long mSize = 0;

    // Number of tuples to sample per bag, to get an estimate of tuple size
    private static final int SPILL_SAMPLE_SIZE = 100;
    private static final int SPILL_SAMPLE_FREQUENCY = 10;

    long aggSampleTupleSize = 0;

    int sampled = 0;

    private boolean spillableRegistered = false;

    /**
     * Get the number of elements in the bag, both in memory and on disk.
     */
    @Override
    public long size() {
        return mSize;
    }


    /**
     * Sample every SPILL_SAMPLE_FREQUENCYth tuple
     * until we reach a max of SPILL_SAMPLE_SIZE
     * to get an estimate of the tuple sizes.
     */
    protected void sampleContents() {
        synchronized (mContents) {
            Iterator<Tuple> iter = mContents.iterator();
            for (int i = 0; i < sampled * SPILL_SAMPLE_FREQUENCY && iter.hasNext(); i++) {
                iter.next();
            }
            for (int i = sampled; iter.hasNext() && sampled < SPILL_SAMPLE_SIZE; i++) {
                Tuple t = iter.next();
                if (t != null && i % SPILL_SAMPLE_FREQUENCY == 0) {
                    aggSampleTupleSize += t.getMemorySize();
                    sampled += 1;
                }
            }
        }
    }

    /**
     * Add a tuple to the bag.
     * @param t tuple to add.
     */
    @Override
    public void add(Tuple t) {
        synchronized (mContents) {
            mSize++;
            mContents.add(t);
        }
        markSpillableIfNecessary();
    }

    /**
     * All bag implementations that can get big enough to be spilled
     * should call this method after every time they add an element.
     */
    protected void markSpillableIfNecessary() {
        if (!spillableRegistered) {
            long estimate = getMemorySize();
            if ( estimate >= SPILL_REGISTER_THRESHOLD) {
                SpillableMemoryManager.getInstance().registerSpillable(this);
                spillableRegistered = true;
            }
        }
    }

    @Override
    public void addAll(DataBag b) {
        addAll((Iterable<Tuple>) b);
        }

    public void addAll(Collection<Tuple> c) {
        addAll((Iterable<Tuple>) c);
    }

    /**
     * Add contents of an iterable (a collection or a DataBag)
     *
     * @param iterable a Collection or DataBag to add contents of
     */
    public void addAll(Iterable<Tuple> iterable) {
        synchronized (mContents) {
            for (Tuple t : iterable) {
                add(t);
            }
        }
    }

    /**
     * Return the size of memory usage.
     */
    @Override
    public long getMemorySize() {
        int numInMem = 0;

        synchronized (mContents) {
            numInMem = mContents.size();

            // If we've already gotten the estimate
            // and the number of tuples hasn't changed, or was above
            // the sample size and is still above the sample size, we can
            // produce a new estimate without sampling the tuples again.
            if (sampled != 0 && (sampled == numInMem ||
                    sampled > SPILL_SAMPLE_SIZE && numInMem > SPILL_SAMPLE_SIZE)) {
                return totalSizeFromAvgTupleSize(aggSampleTupleSize/sampled, numInMem);
            }
            sampleContents();
            int avgTupleSize;
            if (sampled != 0) {
                avgTupleSize = (int) (aggSampleTupleSize / sampled);
            } else {
                avgTupleSize = 0;
            }

           return totalSizeFromAvgTupleSize(avgTupleSize, numInMem);
        }
    }

    private long totalSizeFromAvgTupleSize(long avgTupleSize, int numInMem) {
        long used = avgTupleSize * numInMem;

        long mFields_size =   roundToEight(4 + numInMem*4); /* mContents fixed + per entry */
        // in java hotspot 32bit vm, there seems to be a minimum bag size of 188 bytes
        // some of the extra bytes is probably from a minimum size of this array list
        mFields_size = Math.max(40, mFields_size);

        // the fixed overhead for this object and other object variables = 84 bytes
        // 8 - object header
        // 4 + 8 + 8 - sampled + aggSampleTupleSize + mSize
        // 8 + 8 - mContents ref  + mSpillFiles ref
        // 4 - spillableRegistered +4 instead of 1 to round it to eight
        // 36 - mContents fixed
        used += 84 + mFields_size;

        // add up overhead for mSpillFiles ArrayList, Object[] inside ArrayList,
        // object variable inside ArrayList and references to spill files
        if (mSpillFiles != null) {
            used += roundToEight(36 /* mSpillFiles fixed overhead*/ + mSpillFiles.size()*4);

            if(mSpillFiles.size() > 0){
                //a rough estimate of memory used by each file entry
                // the auto generated files are likely to have same length
                long approx_per_entry_size =
                    roundToEight(mSpillFiles.get(0).toString().length() * 2 + 38);

                used += mSpillFiles.size() * approx_per_entry_size;
            }
        }
        return used;

    }

    /**
     * Memory size of objects are rounded to multiple of 8 bytes
     * @param i
     * @return i rounded to a equal of higher multiple of 8
     */
    private long roundToEight(long i) {
        return 8 * ((i+7)/8); // integer division rounds the result down
    }

    /**
     * Clear out the contents of the bag, both on disk and in memory.
     * Any attempts to read after this is called will produce undefined
     * results.
     */
    @Override
    public void clear() {
        synchronized (mContents) {
            mContents.clear();
            if (mSpillFiles != null) {
                for (int i = 0; i < mSpillFiles.size(); i++) {
                    boolean res = mSpillFiles.get(i).delete();
                    if (!res)
                        warn ("DefaultAbstractBag.clear: failed to delete " + mSpillFiles.get(i), PigWarning.DELETE_FAILED, null);
                }
                mSpillFiles.clear();
            }
            mSize = 0;
            aggSampleTupleSize = 0;
            sampled = 0;
            // not changing spillableRegistered -- clear doesn't change that.
        }
    }

    /**
     * This method is potentially very expensive since it may require a
     * sort of the bag; don't call it unless you have to.
     */
    @Override
    @SuppressWarnings("unchecked")
    public int compareTo(Object other) {
        if (this == other)
            return 0;
        if (other instanceof DataBag) {
            DataBag bOther = (DataBag) other;
            if (this.size() != bOther.size()) {
                if (this.size() > bOther.size()) return 1;
                else return -1;
            }

            // Ugh, this is bogus.  But I have to know if two bags have the
            // same tuples, regardless of order.  Hopefully most of the
            // time the size check above will prevent this.
            // If either bag isn't already sorted, create a sorted bag out
            // of it so I can guarantee order.
            DataBag thisClone;
            DataBag otherClone;
            BagFactory factory = BagFactory.getInstance();

            if (this.isSorted() || this.isDistinct()) {
                thisClone = this;
            } else {
                thisClone = factory.newSortedBag(null);
                Iterator<Tuple> i = iterator();
                while (i.hasNext()) thisClone.add(i.next());

            }
            if (((DataBag) other).isSorted() || ((DataBag)other).isDistinct()) {
                otherClone = bOther;
            } else {
                otherClone = factory.newSortedBag(null);
                Iterator<Tuple> i = bOther.iterator();
                while (i.hasNext()) otherClone.add(i.next());
            }
            Iterator<Tuple> thisIt = thisClone.iterator();
            Iterator<Tuple> otherIt = otherClone.iterator();
            while (thisIt.hasNext() && otherIt.hasNext()) {
                Tuple thisT = thisIt.next();
                Tuple otherT = otherIt.next();

                int c = thisT.compareTo(otherT);
                if (c != 0) return c;
            }

            return 0;   // if we got this far, they must be equal
        } else {
            return DataType.compare(this, other);
        }
    }

    @Override
    public boolean equals(Object other) {
        return compareTo(other) == 0;
    }

    /**
     * Write a bag's contents to disk.
     * @param out DataOutput to write data to.
     * @throws IOException (passes it on from underlying calls).
     */
    @Override
    public void write(DataOutput out) throws IOException {
        sedes.writeDatum(out, this);
    }

    /**
     * Read a bag from disk.
     * @param in DataInput to read data from.
     * @throws IOException (passes it on from underlying calls).
     */
    @Override
    public void readFields(DataInput in) throws IOException {
        long size = in.readLong();

        for (long i = 0; i < size; i++) {
            try {
                Object o = sedes.readDatum(in);
                add((Tuple)o);
            } catch (ExecException ee) {
                throw ee;
            }
        }
    }

    /**
     * This is used by FuncEvalSpec.FakeDataBag.
     * @param stale Set stale state.
     */
    @Override
    public void markStale(boolean stale)
    {
    }

    /**
     * Write the bag into a string. */
    @Override
    public String toString() {
        return BagFormat.format(this);
    }

    @Override
    public int hashCode() {
        int hash = 1;
        Iterator<Tuple> i = iterator();
        while (i.hasNext()) {
            // Use 37 because we want a prime, and tuple uses 31.
            hash = 37 * hash + i.next().hashCode();
        }
        return hash;
    }

    /**
     * Get a file to spill contents to.  The file will be registered in the
     * mSpillFiles array.
     * @return stream to write tuples to.
     */
    protected DataOutputStream getSpillFile() throws IOException {
        if (mSpillFiles == null) {
            // We want to keep the list as small as possible.
            mSpillFiles = new FileList(1);
        }

        String tmpDirName= System.getProperties().getProperty("java.io.tmpdir") ;
        File tmpDir = new File(tmpDirName);

        // if the directory does not exist, create it.
        if (!tmpDir.exists()){
            log.info("Temporary directory doesn't exists. Trying to create: " + tmpDir.getAbsolutePath());
          // Create the directory and see if it was successful
          if (tmpDir.mkdir()){
            log.info("Successfully created temporary directory: " + tmpDir.getAbsolutePath());
          } else {
              // If execution reaches here, it means that we needed to create the directory but
              // were not successful in doing so.
              //
              // If this directory is created recently then we can simply
              // skip creation. This is to address a rare issue occuring in a cluster despite the
              // the fact that spill() makes call to getSpillFile() in a synchronized
              // block.
              if (tmpDir.exists()) {
                log.info("Temporary directory already exists: " + tmpDir.getAbsolutePath());
              } else {
                int errCode = 2111;
                String msg = "Unable to create temporary directory: " + tmpDir.getAbsolutePath();
                throw new ExecException(msg, errCode, PigException.BUG);
              }
          }
        }

        File f = File.createTempFile("pigbag", null);
        f.deleteOnExit();
        mSpillFiles.add(f);
        return new DataOutputStream(new BufferedOutputStream(
            new FileOutputStream(f)));
    }

    /**
     * Report progress to HDFS.
     */
    protected void reportProgress() {
        if (PhysicalOperator.getReporter() != null) {
            PhysicalOperator.getReporter().progress();
        }
    }

    @SuppressWarnings("rawtypes")
    protected void warn(String msg, Enum warningEnum, Exception e) {
        pigLogger = PhysicalOperator.getPigLogger();
        if(pigLogger != null) {
            pigLogger.warn(this, msg, warningEnum);
        } else {
            log.warn(msg, e);
        }
    }

    @SuppressWarnings("rawtypes")
    protected void incSpillCount(Enum counter) {
        incSpillCount(counter, 1);
    }

    @SuppressWarnings("rawtypes")
    protected void incSpillCount(Enum counter, long numRecsSpilled) {
        PigStatusReporter reporter = PigStatusReporter.getInstance();
        if (reporter != null && reporter.getCounter(counter)!=null) {
            reporter.getCounter(counter).increment(numRecsSpilled);
        } else {
            PigHadoopLogger.getInstance().warn(this, "Spill counter incremented", counter);
        }
    }

    public static abstract class BagDelimiterTuple extends DefaultTuple{}
    public static class StartBag extends BagDelimiterTuple{
        private static final long serialVersionUID = 1L;}

    public static class EndBag extends BagDelimiterTuple{
        private static final long serialVersionUID = 1L;}

    public static final Tuple startBag = new StartBag();
    public static final Tuple endBag = new EndBag();

    protected static final int MAX_SPILL_FILES = 100;

}
TOP

Related Classes of org.apache.pig.data.DefaultAbstractBag$EndBag

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.