mac安装Pyspark并连接Mysql

news2025/2/21 11:27:23

安装Scala, apache-spark, Hadoop

brew install scala
brew install apache-spark
brew install hadoop
pip install pyspark

注意不要自己另外安装jdk, 会造成版本对不上报错。因为安装apache-spark的过程中会自动安装openjdk。

配置环境变量

JAVA_HOME=/opt/homebrew/Cellar/openjdk@11/11.0.26/libexec/openjdk.jdk/Contents/Home
export PATH=$PATH:$JAVA_HOME/bin
export CLASSPATH=$JAVA_HOME/lib
export JAVA_HOME

SCALA_HOME=/opt/homebrew/opt/scala
export PATH=$PATH:$SCALA_HOME/bin

export SPARK_PATH=/opt/homebrew/Cellar/apache-spark/3.5.4
export PATH=$PATH:$SPARK_PATH/bin

export openjdk_home=/opt/homebrew/Cellar/openjdk@17/17.0.14


HADOOP_HOME=/opt/homebrew/Cellar/hadoop/3.4.1
export PATH=$PATH:/$HADOOP_HOME/bin

注意jdk版本不对会造成Py4JJavaError: An error occurred while calling None.org.apache.spark.api.java.JavaSparkContext.: java.lang.UnsupportedOperationException: getSubject is supported only if a security manager is allowed

下载jdbc driver
链接, 选择 Platform Independent

将下载的解压包中的jar文件拷贝到apache-spark的目录下: /opt/homebrew/Cellar/apache-spark/3.5.4/libexec/jars

使用如下代码来测试

from pyspark.sql import SparkSession
from pyspark import SparkContext

# sc = SparkContext.getOrCreate()

sparkdriver = SparkSession.builder.\
    appName('demo').\
    master('local[*]').\
    config('spark.driver.extraClassPath', '/opt/homebrew/Cellar/apache-spark/3.5.4/libexec/jars/mysql-connector-j-9.2.0.jar').\
    getOrCreate()
    
df_mysql = sparkdriver.read.format('jdbc').\
    option('url', 'jdbc:mysql://localhost:3306').\
    option('driver', 'com.mysql.jdbc.Driver').\
    option('user', 'root').\
    option('password', '123').\
    option('query', 'select * from tablename').\
    load()

df_mysql.show(10)