Package org.apache.accumulo.core.file.rfile

Interface MetricsGatherer<T>

Type Parameters:

T - Type used to return metrics in getMetrics(). This does not impact collection of metrics at all, is only used in that method.

All Known Implementing Classes:

VisMetricsGatherer

public interface MetricsGatherer<T>

Interface used to gather metrics from RFiles.

Method Summary

Modifier and Type	Method	Description
void	addMetric(Key key, Value val)	Collect and store metrics for the given entry.
T	getMetrics()
void	init(Map<String,ArrayList<ByteSequence>> cf)	Initialize the gatherer when it is registered with the RFile Reader
void	printMetrics(boolean hash, String metricWord, PrintStream out)	Print the results of the metrics gathering by locality group in the format: Metric name Number of keys Percentage of keys Number of blocks Percentage of blocks
void	startBlock()	Start a new block within a LocalityGroup.
void	startLocalityGroup(org.apache.hadoop.io.Text cf)	Start a new LocalityGroup.

Method Detail

init

void init(Map<String,ArrayList<ByteSequence>> cf)

Initialize the gatherer when it is registered with the RFile Reader

Parameters:

cf - Map of the LocalityGroup names to their column families

startLocalityGroup

void startLocalityGroup(org.apache.hadoop.io.Text cf)

Start a new LocalityGroup. This method is used when the RFile seeks to the next LocalityGroup.

Parameters:

cf - Text object of the column family of the first entry in the locality group

addMetric

void addMetric(Key key,
               Value val)

Collect and store metrics for the given entry.

Parameters:

key - Key object of the entry you are collecting metrics from

val - Value object of the entry you are collecting metrics from

startBlock

void startBlock()

Start a new block within a LocalityGroup. This method is used when the RFile moves on the the next block in the LocalityGroup.

printMetrics

void printMetrics(boolean hash,
                  String metricWord,
                  PrintStream out)

Print the results of the metrics gathering by locality group in the format: Metric name Number of keys Percentage of keys Number of blocks Percentage of blocks

Parameters:

hash - Boolean to determine whether the values being printed should be hashed

metricWord - String of the name of the metric that was collected

out - PrintStream of where the information should be written to

getMetrics

T getMetrics()

Returns:

the metrics gathered