

本文属于机器翻译版本。若本译文内容与英语原文存在差异，则一律以英文原文为准。

# 各 Amazon EMR 4.x 发行版之间的差异
<a name="emr-release-differences-4x"></a>

《Amazon EMR 管理指南》**中有关 Amazon EMR 功能的文档指定开始提供某种功能的 Amazon EMR 发行版以及追溯至 4.0.0 的 Amazon EMR 功能之间的适用差异。

从 Amazon EMR 发行版 5.0.0 开始，一些应用程序已进行重大版本升级，安装或运行详细信息改变，另一些应用程序则从沙盒应用程序提升为本机应用程序。此节中的各个主题提供使用 Amazon EMR 4.x 发行版时特定于应用程序的明显差异。

**Topics**
+ [沙盒应用程序](emr-sandbox-apps-4x.md)
+ [使用 Amazon EMR 4.x 上的 Hive 的注意事项](emr-Hive-4x.md)
+ [使用 Amazon EMR 4.x 上的 Pig 的注意事项](emr-Pig-4x.md)

# 沙盒应用程序
<a name="emr-sandbox-apps-4x"></a>

使用 Amazon EMR 4.x 发行版时，一些应用程序会被视为*沙盒*应用程序。沙盒应用程序是我们在初始 Amazon EMR 发行版时因需求提供的应用程序的早期版本。您可以使用控制台或 API 让 Amazon EMR 以与原生应用程序相同的方式安装沙盒应用程序，但是沙盒应用程序的支持和文档有限。 AWS CLI沙盒应用程序在 Amazon EMR 发行版 5.0.0 及更高版本中变为完全受支持的本机应用程序。以下是 Amazon EMR 4.x 发行版中的沙盒应用程序。
+ Oozie
+ ：Presto
+ Sqoop
+ Zeppelin
+ ZooKeeper

当您安装沙盒应用程序时，使用后缀 `-sandbox` 指示应用程序名称。例如，要安装的沙盒版本*Presto*，请使用`Presto-sandbox`。与完全受支持的应用程序相比，安装时间可能更长。此节中列出的每个应用程序的版本号对应于应用程序的社区版本。

## Oozie（沙盒版本）
<a name="emr-Oozie-sandbox-4x"></a>

Oozie 从 Amazon EMR 发行版 4.1.0 开始作为沙盒应用程序提供。

默认情况下，不会使用沙盒版本安装 Oozie 示例。要安装以上示例，可使用 SSH 连接到主节点并运行 `install-oozie-examples`。


**Oozie 沙盒版本信息**  

| Amazon EMR 发行版标签 | Oozie 沙盒版本 | 随 Oozie 沙盒安装的组件 | 
| --- | --- | --- | 
| emr-4.9.6 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.9.5 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.9.4 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.9.3 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.9.2 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.9.1 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.8.5 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.8.4 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.8.3 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.8.2 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.8.1 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.8.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.7.4 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.7.3 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.7.2 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.7.1 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.7.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.6.1 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.6.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.5.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.4.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.3.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.2.0 | 4.2.0 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 
| emr-4.1.0 | 4.0.1 | emrfs, emr-ddb, emr-goodies, emr-kinesis, emr-s3-dist-cp, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, oozie-client, oozie-server | 

## Presto（沙盒版本）
<a name="emr-Presto-sandbox-4x"></a>

Presto 从 Amazon EMR 发行版 4.1.0 开始作为沙盒应用程序提供。


**Presto 沙盒版本信息**  

| Amazon EMR 发行版标签 | Presto 沙盒版本 | 随 Presto 沙盒安装的组件 | 
| --- | --- | --- | 
| emr-4.9.6 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.9.5 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.9.4 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.9.3 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.9.2 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.9.1 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.8.5 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.8.4 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.8.3 | 0.157.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.8.2 | 0.152.3 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.8.1 | 0.151 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.8.0 | 0.151 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.7.4 | 0.148 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.7.3 | 0.148 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.7.2 | 0.148 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.7.1 | 0.147 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.7.0 | 0.147 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hcatalog-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.6.1 | 0.143 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.6.0 | 0.143 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.5.0 | 0.140 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.4.0 | 0.136 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.3.0 | 0.130 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.2.0 | 0.125 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 
| emr-4.1.0 | 0.119 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, hive-client, hive-metastore-server, mysql-server, presto-coordinator, presto-worker | 

## Sqoop（沙盒版本）
<a name="emr-Sqoop-sandbox-4x"></a>

Sqoop 从 Amazon EMR 发行版 4.4.0 开始作为沙盒应用程序提供。


**Sqoop 沙盒版本信息**  

| Amazon EMR 发行版标签 | Sqoop 沙盒版本 | 随 Sqoop 沙盒安装的组件 | 
| --- | --- | --- | 
| emr-4.9.6 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.9.5 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.9.4 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.9.3 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.9.2 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.9.1 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.8.5 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.8.4 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.8.3 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.8.2 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.8.1 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.8.0 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.7.4 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.7.3 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.7.2 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.7.1 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.7.0 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.6.1 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.6.0 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, mysql-server, sqoop-client | 
| emr-4.5.0 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, sqoop-client | 
| emr-4.4.0 | 1.4.6 | emrfs, emr-ddb, emr-goodies, hadoop-client, hadoop-mapred, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, sqoop-client | 

## Zeppelin（沙盒版本）
<a name="emr-Zeppelin-sandbox-4x"></a>

Zeppelin 从 Amazon EMR 发行版 4.1.0 开始作为沙盒应用程序提供。


**Zeppelin 沙盒版本信息**  

| Amazon EMR 发行版标签 | Zeppelin 沙盒版本 | 随 Zeppelin 沙盒安装的组件 | 
| --- | --- | --- | 
| emr-4.9.6 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.9.5 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.9.4 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.9.3 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.9.2 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.9.1 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.8.5 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.8.4 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.8.3 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.8.2 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.8.1 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.8.0 | 0.6.1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.7.4 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.7.3 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.7.2 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.7.1 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.7.0 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.6.1 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.6.0 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.5.0 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.4.0 | 0.5.6-incubating | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.3.0 | 0.5.5-incubating-amzn-1 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.2.0 | 0.5.5-incubating-amzn-0 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 
| emr-4.1.0 | 0.6.0-incubating-SNAPSHOT | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, spark-client, spark-history-server, spark-on-yarn, spark-yarn-slave, zeppelin-server | 

## ZooKeeper （沙盒版本）
<a name="emr-Zookeeper-sandbox-4x"></a>

Zookeeper 从 Amazon EMR 发行版 4.6.0 开始作为沙盒应用程序提供。


**ZooKeeper-沙盒版本信息**  

| Amazon EMR 发行版标签 | ZooKeeper-沙盒版 | 与沙盒一起安装的 ZooKeeper组件 | 
| --- | --- | --- | 
| emr-4.9.6 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.9.5 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.9.4 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.9.3 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.9.2 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.9.1 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.8.5 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.8.4 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.8.3 | 3.4.9 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.8.2 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.8.1 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.8.0 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.7.4 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.7.3 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.7.2 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.7.1 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.7.0 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-kms-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.6.1 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 
| emr-4.6.0 | 3.4.8 | emrfs, emr-goodies, hadoop-client, hadoop-hdfs-datanode, hadoop-hdfs-library, hadoop-hdfs-namenode, hadoop-httpfs-server, hadoop-yarn-nodemanager, hadoop-yarn-resourcemanager, zookeeper-client, zookeeper-server | 

# 使用 Amazon EMR 4.x 上的 Hive 的注意事项
<a name="emr-Hive-4x"></a>

本节介绍在使用 Amazon EMR 4.x 发行版上的 Hive 版本 1.0.0 时要注意的差异（与 Amazon EMR 5.x 发行版上的 Hive 2.x 相比）。

## 不支持 ACID 事务
<a name="emr-Hive-acid-4x"></a>

使用 Amazon EMR 4.x 发行版时，4.x 发行版上的 Hive 不支持 Hive 数据存储在 Amazon S3 中的 ACID 事务。如果您尝试在 Amazon S3 中创建事务表，将出现异常。

## 对 Amazon S3 中的表的读写操作
<a name="emr-Hive-s3table-4x"></a>

Amazon EMR 4.x 发行版上的 Hive 可直接写入 Amazon S3，无需使用临时文件。这可以改善性能，但导致您无法在同一个 Hive 语句内读写 Amazon S3 中的同一个表。一种解决办法是在 HDFS 中创建并使用临时表。

以下示例显示了如何使用多 Hive 语句更新 Amazon S3 中的表。这些语句基于 Amazon S3 中一个名为 `my_s3_table` 的表在 HDFS 中创建一个名为 `tmp` 的临时表。之后将使用此临时表的内容更新 Amazon S3 中的表。

```
CREATE TEMPORARY TABLE tmp LIKE my_s3_table;
INSERT OVERWRITE TABLE tmp SELECT ....;
INSERT OVERWRITE TABLE my_s3_table SELECT * FROM tmp;
```

## Log4j 与 Log4j 2 的对比
<a name="emr-Hive-log4j-4x"></a>

Amazon EMR 4.x 发行版上的 Hive 使用 Log4j。从版本 5.0.0 开始，默认情况下使用 Log4j 2。这些版本可能需要不同的日志记录配置。有关更多信息，请参阅 [Apache Log4j 2](http://logging.apache.org/log4j/2.x/)。

## MapReduce 是默认的执行引擎
<a name="emr-Hive-tez-4x"></a>

亚马逊 EMR 4.x 发行版上的 Hive MapReduce 用作默认执行引擎。从 Amazon EMR 版本 5.0.0 开始，Tez 为默认引擎，这将改善大多数工作流的性能。

## Hive 授权
<a name="emr-Hive-authz-4x"></a>

Amazon EMR 4.x 发行版上的 Hive 对于 HDFS 支持 [Hive 授权](https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Authorization)，但对于 EMRFS 和 Amazon S3 不支持此授权。默认情况下，Amazon EMR 集群在禁用授权的状态下运行。

## Amazon S3 中的 Hive 文件合并操作
<a name="emr-Hive-filemerge-4x"></a>

如果 `hive.merge.mapfiles` 为 `true`，Amazon EMR 4.x 发行版上的 Hive 将在仅映射任务结束时合并小型文件。仅当任务的平均输出大小低于 `hive.merge.smallfiles.avgsize` 设置时，才会触发合并。如果最终输出路径位于 HDFS 中，那么 Amazon EMR Hive 的行为将完全相同。但是，如果输出路径位于 Amazon S3 中，将忽略 `hive.merge.smallfiles.avgsize` 参数。在那种情况下，如果 `hive.merge.mapfiles` 设置为 `true`，会始终触发合并任务。

# 使用 Amazon EMR 4.x 上的 Pig 的注意事项
<a name="emr-Pig-4x"></a>

Pig 版本 0.14.0 安装在使用 Amazon EMR 4.x 发行版创建的集群上。Pig 在 Amazon EMR 5.0.0 中已升级到版本 0.16.0。下面介绍了明显差异。

## 不同的默认执行引擎
<a name="emr-Pig-engine-4x"></a>

亚马逊 EMR 4.x 发行版上的 Pig 版本 0.14.0 MapReduce 用作默认执行引擎。Pig 0.16.0 及更高版本均使用 Apache Tez。你可以在`pig-properties`配置分类`exectype=mapreduce`中明确设置要使用 MapReduce。

## 丢弃了 Pig 用户定义的函数 (UDFs)
<a name="emr-Pig-udf-4x"></a>

Pig on Amazon EMR 4.x 发行版本中提供的自定义 UDFs 版本从 Pig 0.16.0 开始被删除。其中大多数 UDFs 都有等效的函数，你可以改用。下表列出了删除的函数 UDFs 和等效函数。有关更多信息，请参阅 Apache Pig 网站上的[内置函数](https://pig.apache.org/docs/r0.16.0/func.html)。


| 已删除的 UDF | 等效函数 | 
| --- | --- | 
|  FORMAT\$1DT(dtformat, date)  |  GetHour（日期）、 GetMinute（日期）、 GetMonth（日期）、 GetSecond（日期）、 GetWeek（日期）、 GetYear（日期）、 GetDay（日期）  | 
|  EXTRACT(string, pattern)  |  REGEX\$1EXTRACT\$1ALL(string, pattern)  | 
|  REPLACE(string, pattern, replacement)  |  REPLACE(string, pattern, replacement)  | 
|  DATE\$1TIME()  |  ToDate()  | 
|  DURATION(dt, dt2)  |  WeeksBetween(dt, dt2), (dt, dt2), YearsBetween (dt, dt2), SecondsBetween (dt, dt2), (dt, dt2), MonthsBetween (dt, dt2), MinutesBetween (dt, dt2) HoursBetween  | 
|  EXTRACT\$1DT(format, date)  |  GetHour（日期）、 GetMinute（日期）、 GetMonth（日期）、 GetSecond（日期）、 GetWeek（日期）、 GetYear（日期）、 GetDay（日期）  | 
|  OFFSET\$1DT(date, duration)  |  AddDuration（日期、持续时间）、 SubtractDuration（日期、持续时间）  | 
|  PERIOD(dt, dt2)  |  WeeksBetween(dt, dt2), (dt, dt2), YearsBetween (dt, dt2), SecondsBetween (dt, dt2), (dt, dt2), MonthsBetween (dt, dt2), MinutesBetween (dt, dt2) HoursBetween  | 
|  CAPITALIZE(string)  |  UCFIRST(string)  | 
|  CONCAT\$1WITH()  |  CONCAT()  | 
|  INDEX\$1OF()  |  INDEXOF()  | 
|  LAST\$1INDEX\$1OF()  |  LAST\$1INDEXOF()  | 
|  SPLIT\$1ON\$1REGEX()  |  STRSPLT()  | 
|  UNCAPITALIZE()  |  LCFIRST()  | 

以下内容 UDFs 被删除但没有等效项：FORMAT ()、LOCAL\$1DATE ()、LOCAL\$1TIME ()、CENTER ()、LEFT\$1PAD ()、REPLET\$1ONCE ()、RIP\$1PAD ()、STRIP\$1END ()、STRIP\$1END ()、STRIP\$1START ()、SWAP\$1CASE ()。

## 已停止使用 Grunt 命令
<a name="emr-pig-gruntcmd-4x"></a>

某些 Grunt 命令已从 Pig 0.16.0 开始停用。下表列出了 Pig 0.14.0 中的 Grunt 命令以及当前版本中的等效命令（如果适用）。


**Pig 0.14.0 和等效的当前 Grunt 命令**  

| Pig 0.14.0 Grunt 命令 | 0.16.0 及更高版本中的 Pig Grunt 命令 | 
| --- | --- | 
|  cat < non-hdfs-path >)  |  fs-cat < non-hdfs-path >;  | 
| cd < non-hdfs-path >; |  无等效函数  | 
| ls < non-hdfs-path >; | fs-ls < non-hdfs-path >; | 
|  移动 < non-hdfs-path > < non-hdfs-path >;  |  fs-mv < non-hdfs-path > < non-hdfs-path >;  | 
| 复制 < non-hdfs-path > < non-hdfs-path >; |  fs-cp < non-hdfs-path > < non-hdfs-path >;  | 
| copyToLocal < non-hdfs-path ><local-path>; |  fs-copyToLocal < non-hdfs-path ><local-path>;  | 
| copyFromLocal <local-path>< non-hdfs-path >; |  fs-copyFromLocal <local-path>< non-hdfs-path >;  | 
| mkdir < non-hdfs-path >; |  fs-mkdir < >; non-hdfs-path   | 
| rm < non-hdfs-path >; |  fs-rm-r-r-skipTrash < >; non-hdfs-path   | 
|  rmf < non-hdfs-path >;  |  fs-rm-r-r-skipTrash < >; non-hdfs-path   | 

## 针对非 HDFS 主目录删除的功能
<a name="emr-Pig-users-4x"></a>

Amazon EMR 4.x 发行版上的 Pig 0.14.0 具有两种机制，以允许无主目录的 `hadoop` 用户之外的用户运行 Pig 脚本。第一种机制是自动后备，将初始工作目录设置为根目录 (如果主目录不存在)。第二种机制是 `pig.initial.fs.name` 属性，它允许您更改初始工作目录。

这两种机制从 Amazon EMR 版本 5.0.0 开始不可用，因此用户必须在 HDFS 上有一个主目录。这不适用于 `hadoop` 用户，因为在启动时会配置一个主目录。使用 Hadoop jar 步骤运行的脚本默认为由 Hadoop 用户运行，除非使用 `command-runner.jar` 显式指定了其它用户。