org.apache.hadoop.hbase.util

Class CollectionUtils

java.lang.Object

org.apache.hadoop.hbase.util.CollectionUtils

@InterfaceAudience.Private
public class CollectionUtils
extends Object

Utility methods for dealing with Collections, including treating null collections as empty.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static interface	CollectionUtils.IOExceptionSupplier<V> A supplier that throws IOException when get.

Constructor Summary

Constructors

Constructor and Description
CollectionUtils()

Method Summary

Modifier and Type	Method and Description
static <K,V> V	computeIfAbsent(ConcurrentMap<K,V> map, K key, Supplier<V> supplier) In HBASE-16648 we found that ConcurrentHashMap.get is much faster than computeIfAbsent if the value already exists.
static <K,V> V	computeIfAbsent(ConcurrentMap<K,V> map, K key, Supplier<V> supplier, Runnable actionIfAbsent)
static <K,V> V	computeIfAbsentEx(ConcurrentMap<K,V> map, K key, CollectionUtils.IOExceptionSupplier<V> supplier) In HBASE-16648 we found that ConcurrentHashMap.get is much faster than computeIfAbsent if the value already exists.

Methods inherited from class java.lang.Object

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

Constructor Detail

CollectionUtils

public CollectionUtils()

Method Detail

computeIfAbsent

public static <K,V> V computeIfAbsent(ConcurrentMap<K,V> map,
                                      K key,
                                      Supplier<V> supplier)

In HBASE-16648 we found that ConcurrentHashMap.get is much faster than computeIfAbsent if the value already exists. Notice that the implementation does not guarantee that the supplier will only be executed once.

computeIfAbsentEx

public static <K,V> V computeIfAbsentEx(ConcurrentMap<K,V> map,
                                        K key,
                                        CollectionUtils.IOExceptionSupplier<V> supplier)
                                 throws IOException

In HBASE-16648 we found that ConcurrentHashMap.get is much faster than computeIfAbsent if the value already exists. So here we copy the implementation of ConcurrentMap.computeIfAbsent(Object, java.util.function.Function). It uses get and putIfAbsent to implement computeIfAbsent. And notice that the implementation does not guarantee that the supplier will only be executed once.

Throws:

IOException

computeIfAbsent

public static <K,V> V computeIfAbsent(ConcurrentMap<K,V> map,
                                      K key,
                                      Supplier<V> supplier,
                                      Runnable actionIfAbsent)