org.apache.pig.tools.pigstats.mapreduce

Class MRPigStatsUtil

java.lang.Object

org.apache.pig.tools.pigstats.PigStatsUtil

org.apache.pig.tools.pigstats.mapreduce.MRPigStatsUtil

public class MRPigStatsUtil
extends PigStatsUtil

A utility class for Pig Statistics

Field Summary

Fields

Modifier and Type	Field and Description
static String	FS_COUNTER_GROUP
static String	TASK_COUNTER_GROUP

Fields inherited from class org.apache.pig.tools.pigstats.PigStatsUtil

FILE_BYTES_READ, FILE_BYTES_WRITTEN, HDFS_BYTES_READ, HDFS_BYTES_WRITTEN, MAP_INPUT_RECORDS, MAP_OUTPUT_RECORDS, MULTI_INPUTS_COUNTER_GROUP, MULTI_INPUTS_RECORD_COUNTER, MULTI_STORE_COUNTER_GROUP, MULTI_STORE_RECORD_COUNTER, REDUCE_INPUT_RECORDS, REDUCE_OUTPUT_RECORDS

Constructor Summary

Constructors

Constructor and Description
MRPigStatsUtil()

Method Summary

Modifier and Type	Method and Description
static void	accumulateStats(org.apache.hadoop.mapred.jobcontrol.JobControl jc) Updates the statistics after a patch of jobs is done
static JobStats	addJobStats(org.apache.hadoop.mapred.jobcontrol.Job job) Add stats for a new Job, which doesn't yet need to be completed.
static MRJobStats	addNativeJobStats(PigStats ps, NativeMapReduceOper mr, boolean success)
static MRJobStats	addNativeJobStats(PigStats ps, NativeMapReduceOper mr, boolean success, Exception e)
static void	displayStatistics() Logs the statistics in the Pig log file at INFO level
static long	getMultiStoreCount(org.apache.hadoop.mapred.jobcontrol.Job job, org.apache.hadoop.mapred.JobClient jobClient, String counterName) Returns the count for the given counter name in the counter group 'MultiStoreCounters'
static void	setBackendException(org.apache.hadoop.mapred.jobcontrol.Job job, Exception e)
static void	startCollection(PigContext pc, org.apache.hadoop.mapred.JobClient client, JobControlCompiler jcc, MROperPlan plan) Starts collecting statistics for the given MR plan
static void	stopCollection(boolean display) Stops collecting statistics for a MR plan
static void	updateJobMroMap(Map<org.apache.hadoop.mapred.jobcontrol.Job,MapReduceOper> jobMroMap) Updates the PigStats.JobGraph of the PigStats.

Methods inherited from class org.apache.pig.tools.pigstats.PigStatsUtil

getEmptyPigStats, getMultiInputsCounterName, getMultiStoreCounterName, getPigStats, getShortName, isTempFile, setErrorCode, setErrorMessage, setErrorThrowable, setStatsMap

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

TASK_COUNTER_GROUP

public static final String TASK_COUNTER_GROUP

See Also:

Constant Field Values

FS_COUNTER_GROUP

public static final String FS_COUNTER_GROUP

See Also:

Constant Field Values

Constructor Detail

MRPigStatsUtil

public MRPigStatsUtil()

Method Detail

getMultiStoreCount

public static long getMultiStoreCount(org.apache.hadoop.mapred.jobcontrol.Job job,
                                      org.apache.hadoop.mapred.JobClient jobClient,
                                      String counterName)

Returns the count for the given counter name in the counter group 'MultiStoreCounters'

Parameters:

job - the MR job

jobClient - the Hadoop job client

counterName - the counter name

Returns:

the count of the given counter name

startCollection

public static void startCollection(PigContext pc,
                                   org.apache.hadoop.mapred.JobClient client,
                                   JobControlCompiler jcc,
                                   MROperPlan plan)

Starts collecting statistics for the given MR plan

Parameters:

pc - the Pig context

client - the Hadoop job client

jcc - the job compiler

plan - the MR plan

stopCollection

public static void stopCollection(boolean display)

Stops collecting statistics for a MR plan

Parameters:

display - if true, log collected statistics in the Pig log file at INFO level

addJobStats

public static JobStats addJobStats(org.apache.hadoop.mapred.jobcontrol.Job job)

Add stats for a new Job, which doesn't yet need to be completed.

Parameters:

job - the job being run

Returns:

JobStats for the job

displayStatistics

public static void displayStatistics()

Logs the statistics in the Pig log file at INFO level

updateJobMroMap

public static void updateJobMroMap(Map<org.apache.hadoop.mapred.jobcontrol.Job,MapReduceOper> jobMroMap)

Updates the PigStats.JobGraph of the PigStats. The initial PigStats.JobGraph is created without job ids using MROperPlan, before any job is submitted for execution. The PigStats.JobGraph then is updated with job ids after jobs are executed.

Parameters:

jobMroMap - the map that maps Jobs to MapReduceOpers

accumulateStats

public static void accumulateStats(org.apache.hadoop.mapred.jobcontrol.JobControl jc)

Updates the statistics after a patch of jobs is done

Parameters:

jc - the job control

setBackendException

@InterfaceAudience.Private
public static void setBackendException(org.apache.hadoop.mapred.jobcontrol.Job job,
                                                                  Exception e)

addNativeJobStats

public static MRJobStats addNativeJobStats(PigStats ps,
                                           NativeMapReduceOper mr,
                                           boolean success)

addNativeJobStats

public static MRJobStats addNativeJobStats(PigStats ps,
                                           NativeMapReduceOper mr,
                                           boolean success,
                                           Exception e)