pyspark访问hive数据实战

数据分析都是直接使用hive脚本进行调用，随着APP用户行为和日志数据量的逐渐累积，跑每天的脚本运行需要花的时间越来越长，虽然进行了sql优化，但是上spark已经提上日程。

直接进行spark开发需要去学习scala，为了降低数据分析师的学习成本，决定前期先试用sparkSQL，能够让计算引擎无缝从MR切换到spark，现在主要使用pyspark访问hive数据。

以下是安装配置过程中的详细步骤：

1.安装spark

需要先安装JDK和scala，这不必多说，由于现有hadoop集群版本是采用的2.6.3，所以spark版本是下载的稳定版本spark-1.4.0-bin-hadoop2.6.tgz

我是先在一台机器上完成了Spark的部署，Master和Slave都在一台机器上。注意要配置免秘钥ssh登陆。

1.1 环境变量配置

export JAVA_HOME=/usr/jdk1.8.0_73
export HADOOP_HOME=/usr/hadoop
export HADOOP_CONF_DIR=/usr/hadoop/etc/hadoop
export SCALA_HOME=/usr/local/scala-2.11.7
export SPARK_HOME=/home/hadoop/spark_folder/spark-1.4.0-bin-hadoop2.6
export SPARK_MASTER_IP=127.0.0.1
export SPARK_MASTER_PORT=7077
export SPARK_MASTER_WEBUI_PORT=8099
export SPARK_WORKER_CORES=3 //每个Worker使用的CPU核数
export SPARK_WORKER_INSTANCES=1 //每个Slave中启动几个Worker实例
export SPARK_WORKER_MEMORY=10G //每个Worker使用多大的内存
export SPARK_WORKER_WEBUI_PORT=8081 //Worker的WebUI端口号
export SPARK_EXECUTOR_CORES=1 //每个Executor使用使用的核数
export SPARK_EXECUTOR_MEMORY=1G //每个Executor使用的内存
export HIVE_HOME=/home/hadoop/hive
export SPARK_CLASSPATH=$HIVE_HOME/lib/mysql-connector-java-5.1.31-bin.jar:$SPARK_CLASSPATH
export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:$HADOOP_HOME/lib/native

1.2 配置slaves

cp slaves.template slaves
vi slaves 添加以下内容：localhost

1.3 启动master和slave

cd $SPARK_HOME/sbin/
./start-master.sh
启动日志位于 $SPARK_HOME/logs/目录，访问 http://localhost:8099，即可看到Spark的WebUI界面
执行 ./bin/spark-shell,打开Scala到Spark的连接窗口

2.SparkSQL与Hive的整合

拷贝$HIVE_HOME/conf/hive-site.xml和hive-log4j.properties到 $SPARK_HOME/conf/
在$SPARK_HOME/conf/目录中，修改spark-env.sh，添加
export HIVE_HOME=/home/hadoop/hive
export SPARK_CLASSPATH=$HIVE_HOME/lib/mysql-connector-java-5.1.31-bin.jar:$SPARK_CLASSPATH
另外也可以设置一下Spark的log4j配置文件，使得屏幕中不打印额外的INFO信息(如果不想受干扰可设置为更高):
log4j.rootCategory=WARN, console
进入$SPARK_HOME/bin，执行 ./spark-sql –master spark://127.0.0.1:7077 进入spark-sql CLI:
[hadoop@hadoop spark]$ bin/spark-sql --help
Usage: ./bin/spark-sql [options] [cli option]
CLI options:
-d,--define Variable subsitution to apply to hive
commands. e.g. -d A=B or --define A=B
--database Specify the database to use
-e SQL from command line
-f SQL from files
-h connecting to Hive Server on remote host
--hiveconf Use value for given property
--hivevar Variable subsitution to apply to hive
commands. e.g. --hivevar A=B
-i Initialization SQL file
-p connecting to Hive Server on port number
-S,--silent Silent mode in interactive shell
-v,--verbose Verbose mode (echo executed SQL to the
console)

需要注意的是CLI不是使用JDBC连接，所以不能连接到ThriftServer;但可以配置conf/hive-site.xml连接到hive的metastore，然后对hive数据进行查询。下面我们接着说如何在python中连接hive数据表查询。

3.配置pyspark和示例代码

3.1 配置pyspark

打开/etc/profile:
#PythonPath 将Spark中的pySpark模块增加的Python环境中
export PYTHONPATH=/opt/spark-hadoop/python
source /etc/profile

执行./bin/pyspark ,打开Python到Spark的连接窗口，确认没有报错。

打开命令行窗口，输入python,Python版本为2.7.6，如图所示，注意Spark暂时不支持Python3。输入import pyspark不报错，证明开发前工作已经完成。

3.2 启动ThriftServer

启动ThriftServer，使之运行在spark集群中：

sbin/start-thriftserver.sh --master spark://localhost:7077 --executor-memory 5g

ThriftServer可以连接多个JDBC/ODBC客户端，并相互之间可以共享数据。

3.3 请求示例

查看spark官方文档说明，spark1.4和2.0对于sparksql调用hive数据的API变化并不大。都是用sparkContext 。

from pyspark import SparkConf, SparkContext
from pyspark.sql import HiveContext
conf = (SparkConf()
.setMaster("spark://127.0.0.1:7077")
.setAppName("My app")
.set("spark.executor.memory", "1g"))
sc = SparkContext(conf = conf)
sqlContext = HiveContext(sc)
my_dataframe = sqlContext.sql("Select count(1) from logs.fmnews_dim_where")
my_dataframe.show()

返回结果：

运行以后在webUI界面看到job运行详情。

4.性能比较

截取了接近一个月的用户行为数据，数据大小为2G，总共接近1600w条记录。

为了测试不同sql需求情况下的结果，我们选取了日常运行的2类sql：

1.统计数据条数：

select count(1) from fmnews_user_log2;

2.统计用户行为：

SELECT device_id, min_time FROM
(SELECT device_id,min(import_time) min_time FROM fmnews_user_log2
GROUP BY device_id)a
WHERE from_unixtime(int(substr(min_time,0,10)),'yyyy-MM-dd') = '2017-03-02';

3. 用户行为分析：

select case when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '06:00' and '07:59' then 1
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '08:00' and '09:59' then 2
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '10:00' and '11:59' then 3
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '12:00' and '13:59' then 4
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '14:00' and '15:59' then 5
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '16:00' and '17:59' then 6
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '18:00' and '19:59' then 7
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '20:00' and '21:59' then 8
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '22:00' and '23:59' then 9
else 0 end fmnews_time_type, count(distinct device_id) device_count,count(1) click_count
from fmcm.fmnews_user_log2
where from_unixtime(int(substr(import_time,0,10)),'yyyy-MM-dd') = '2017-03-02'
group by case when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '06:00' and '07:59' then 1
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '08:00' and '09:59' then 2
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '10:00' and '11:59' then 3
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '12:00' and '13:59' then 4
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '14:00' and '15:59' then 5
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '16:00' and '17:59' then 6
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '18:00' and '19:59' then 7
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '20:00' and '21:59' then 8
when from_unixtime(int(substr(fmnews_time,0,10)),'HH:mm') between '22:00' and '23:59' then 9
else 0 end;