Flink之PyFlink入门：macOS源码安装flink并提交入门demo

1、克隆flink源码

git clone https://github.com/apache/flink

2、maven构建flink包，留意下方对jdk版本的要求：

In addition you need Maven 3 and a JDK (Java Development Kit). Flink requires at least Java 8 to build.

mvn clean install -DskipTests 
cd flink-dist mvn clean install

若无maven需要安装maven：macOS：brew install maven；ubuntu：apt-get install maven；centos：yum -y install maven

3、构建完成后flink根目录下，构建PyFlink的源码发布包和wheel包：

cd flink-python; python3 setup.py sdist bdist_wheel

4、构建好的源码发布包和wheel包位于./flink-python/dist/目录下。它们均可使用pip安装,比如:

pip install dist/*.tar.gz

5、因为安装了anaconda，pip将pyflink安装到了anaonda默认的路径下：

cd /anaconda3/lib/python3.7/site-packages/pyflink/

6、启动flink集群模式：

sudo bin/start-cluster.sh

启动后可在http://127.0.0.1:8081 查看flink webui

7、也可启动flink的Python交互式界面：

sudo bin/pyflink-shell.sh local

8、若想更改默认的Python交互工具为ipython, 简单修改启动脚本该行即可：

vim bin/pyflink-shell.sh 

PYFLINK_PYTHON="${PYFLINK_PYTHON:-"python"}" 
——> 
PYFLINK_PYTHON="${PYFLINK_PYTHON:-"ipython"}"

注意：需要预先安装ipython

sudo pip3 install ipython -i [https://pypi.tuna.tsinghua.edu.cn/simple](https://pypi.tuna.tsinghua.edu.cn/simple)

9、maven构建的flink根目录下有几个flink的练习demo：

cd ~/flink/flink-examples/

练习demo

10、练习1：启动wordcount.py :

sudo bin/flink run -py examples/python/table/batch/word_count.py

Flink wordcount

wordcount 结果

11、练习2：提交jar包任务：

sudo bin/flink run -j flink-examples-batch/target/flink-examples-batch_2.11-1.10-SNAPSHOT-KMeans.jar

Flink KMeans