org.apache.flink.optimizer.dataproperties

枚举 PartitioningProperty

java.lang.Object

java.lang.Enum<PartitioningProperty>

org.apache.flink.optimizer.dataproperties.PartitioningProperty

所有已实现的接口:

Serializable, Comparable<PartitioningProperty>

public enum PartitioningProperty
extends Enum<PartitioningProperty>

An enumeration of the different types of distributing data across partitions or parallel workers.

枚举常量概要

枚举常量

枚举常量和说明
ANY_DISTRIBUTION Any possible way of data distribution, including random partitioning and full replication.
ANY_PARTITIONING A not further specified partitioning on a key (hash-, or range partitioning, or some other scheme even).
CUSTOM_PARTITIONING A custom partitioning, accompanied by a Partitioner.
FORCED_REBALANCED A forced even re-balancing.
FULL_REPLICATION Full replication of the data to each parallel instance.
HASH_PARTITIONED A hash partitioning on a certain key.
RANDOM_PARTITIONED A random disjunct (non-replicated) data distribution, where each datum is contained in one partition only.
RANGE_PARTITIONED A range partitioning on a certain key.

方法概要

限定符和类型	方法和说明
boolean	isComputablyPartitioned() Checks, if this property represents a partitioning that is computable.
boolean	isPartitioned() Checks, if this property represents in fact a partitioning.
boolean	isPartitionedOnKey() Checks if this property presents a partitioning that is not random, but on a partitioning key.
boolean	isReplication() Checks, if this property represents a full replication.
static PartitioningProperty	valueOf(String name) 返回带有指定名称的该类型的枚举常量。
static PartitioningProperty[]	values() 按照声明该枚举类型的常量的顺序, 返回 包含这些常量的数组。

从类继承的方法 java.lang.Enum

clone, compareTo, equals, finalize, getDeclaringClass, hashCode, name, ordinal, toString, valueOf

从类继承的方法 java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

枚举常量详细资料

ANY_DISTRIBUTION

public static final PartitioningProperty ANY_DISTRIBUTION

Any possible way of data distribution, including random partitioning and full replication.

RANDOM_PARTITIONED

public static final PartitioningProperty RANDOM_PARTITIONED

A random disjunct (non-replicated) data distribution, where each datum is contained in one partition only. This is for example the result of parallel scans of data in a file system like HDFS, or the result of a round-robin data distribution.

HASH_PARTITIONED

public static final PartitioningProperty HASH_PARTITIONED

A hash partitioning on a certain key.

RANGE_PARTITIONED

public static final PartitioningProperty RANGE_PARTITIONED

A range partitioning on a certain key.

ANY_PARTITIONING

public static final PartitioningProperty ANY_PARTITIONING

A not further specified partitioning on a key (hash-, or range partitioning, or some other scheme even).

FULL_REPLICATION

public static final PartitioningProperty FULL_REPLICATION

Full replication of the data to each parallel instance.

FORCED_REBALANCED

public static final PartitioningProperty FORCED_REBALANCED

A forced even re-balancing. All partitions are guaranteed to have almost the same number of records.

CUSTOM_PARTITIONING

public static final PartitioningProperty CUSTOM_PARTITIONING

A custom partitioning, accompanied by a Partitioner.

方法详细资料

values

public static PartitioningProperty[] values()

按照声明该枚举类型的常量的顺序, 返回 包含这些常量的数组。该方法可用于迭代 常量, 如下所示:

for (PartitioningProperty c : PartitioningProperty.values())
    System.out.println(c);

返回:

按照声明该枚举类型的常量的顺序返回的包含这些常量的数组

valueOf

public static PartitioningProperty valueOf(String name)

返回带有指定名称的该类型的枚举常量。 字符串必须与用于声明该类型的枚举常量的 标识符完全匹配。(不允许有多余 的空格字符。)

参数:

name - 要返回的枚举常量的名称。

返回:

返回带有指定名称的枚举常量

抛出:

IllegalArgumentException - 如果该枚举类型没有带有指定名称的常量

NullPointerException - 如果参数为空值

isPartitioned

public boolean isPartitioned()

Checks, if this property represents in fact a partitioning. That is, whether this property is not equal to PartitionProperty.FULL_REPLICATION.

返回:

True, if this enum constant is unequal to PartitionProperty.FULL_REPLICATION, false otherwise.

isReplication

public boolean isReplication()

Checks, if this property represents a full replication.

返回:

True, if this enum constant is equal to PartitionProperty.FULL_REPLICATION, false otherwise.

isPartitionedOnKey

public boolean isPartitionedOnKey()

Checks if this property presents a partitioning that is not random, but on a partitioning key.

返回:

True, if the data is partitioned on a key.

isComputablyPartitioned

public boolean isComputablyPartitioned()

Checks, if this property represents a partitioning that is computable. A computable partitioning can be recreated through an algorithm. If two sets of data are to be co-partitioned, it is crucial, that the partitioning schemes are computable.

Examples for computable partitioning schemes are hash- or range-partitioning. An example for a non-computable partitioning is the implicit partitioning that exists though a globally unique key.

返回:

True, if this enum constant is a re-computable partitioning.