ApacheFlinkv1.9官方中文文档

Apache Flink是一个分布式流批一体化的开源平台。Flink的核心是一个提供数据分发、通信以及自动容错的流计算引擎。Flink在流计算之上构建批

处理，并且原生的支持迭代计算，内存管理以及程序优化。

DataStream API 教程

In this guide we will start from scratch and go from setting up a Flink project to runninga streaming analysis program on a Flink cluster.

wikipedia provides an IRC channel where all edits to the wiki are logged.we are going toread this channel in Flink and count the number of bytes that each user edits withina given window of time.This is easy enough to implement in a few minutes using Flink,but it willgive you a good foundation from which to start building more complex analysis programs on your own.

Setting up a Maven Project

We are going to use a Flink Maven Archetype for creating our project structure. Pleasesee java API Quickstart for more detailsabout this.For our purposes,the command to run is this:

You can edit the grouprd ,artifactId and package if you like.With the above parameters,Maven will create a project structure that looks like this:

There is our pom.xml file that already has the Flink dependencies added in the root directory andseveral example Flink programs in src/main/java.we can delete the example programs,sincewe are going to start from scratch:

As a last step we need to add the Flink wikipedia connector as a dependency so that we canuse it in our program. Edit the dependencies section of the pom.xml so that it looks like this :

Notice the flink-connector-wikiedits2.11 dependency that was added.(This example andthe wikipedia connector were inspired by the _Hello Samza example of Apache Samza.)

Writing a Flink Program

It's coding time.Fire up your favorite IDE and import the Maven project or open a text editor andcreate the file src/main/java/wikiedits/wikipediaAnalysis.java:

The program is very basic now,but we will fill it in as we go. Note that I'll not giveimport statements here since IDEs can add them automatically.At the end of this section I'll showthe complete code with import statements if you simply want to skip ahead and enter that in youreditor.

The first step in a Flink program is to Create a streamExecutionEnvironment (or ExecutionEnvironment if you are writing a batch job). This can be used to set executionparameters and create sources for reading from external systems.So let's go ahead and addthis to the main method:

【下载地址】

百度网盘链接：https://pan.baidu.com/s/1mKn5gwx3eSDLoagsbrNG5g

提取码：r5vt