Python学习从0到1 day26 第三阶段 Spark ④ 数据输出
半山腰太挤了,你该去山顶看看
—— 24.11.10
一、输出为python对象
1.collect算子
功能:
将RDD各个分区内的数据,统一收集到Driver中,形成一个List对象
语法:
rdd.collect()
返回值是一个list列表
示例:
from pyspark import SparkConf,SparkContext
import os
conf = SparkConf().setMaster("local").setAppName("test_spark")
os.environ['PYSPARK_PYTHON'] = "E:/python.learning/pyt/scripts/python.exe"
sc = SparkContext(conf = conf)
Set = {"小明","小红","小强"}
Tuple = ("小明","小红","小强")
set_rdd = sc.parallelize(Set)
tuple_rdd = sc.parallelize(Tuple)
print(set_rdd.collect())
print(tuple_rdd.collect())
2.reduce算子
功能:
对RDD数据集按照你传入的逻辑进行聚合
语法:
rdd.reduce(func)
rdd = sc.parallelize(range(1 , 10))
# 将rdd的数据进行累加求和
print(rdd.reduce(lambda a , b : a + b))
返回值等同于计算函数的返回值
示例:
from pyspark import SparkContext,SparkConf
import os
import json
os.environ['PYSPARK_PYTHON'] = "E:/python.learning/pyt/scripts/python.exe"
conf = SparkConf().setMaster("local").setAppName("test_spark")
sc = SparkContext(conf = conf)
List = [1,2,3,4,5,6,7,8,9]
rdd = sc.parallelize(List)
print(rdd.reduce(lambda x, y : x + y))
3.take算子
功能:
取RDD的前N个元素,组合成list返回
语法:
sc.parallelize([3,2,1,4,5,6]).take(5) # [3,2,1,4,5]
返回前n个元素组成的list
示例:
from pyspark import SparkContext,SparkConf
import os
import json
os.environ['PYSPARK_PYTHON'] = "E:/python.learning/pyt/scripts/python.exe"
conf = SparkConf().setMaster("local[*]").setAppName("test_spark")
sc = SparkContext(conf=conf)
List = (1,2,3,4,5,6,7,8,9)
rdd = sc.parallelize(List)
res = rdd.take(4)
print("前四个元素为:"+res)
4.count算子
功能:
计算RDD有多少条数据
语法:
sc.parallelize([3,2,1,4,5,6]).count()
返回值是一个数字
示例:
from pyspark import SparkConf,SparkContext
import os
import json
os.environ['PYSPARK_PYTHON'] = "E:/python.learning/pyt/scripts/python.exe"
conf = SparkConf().setMaster("local[*]").setAppName("test_spark")
sc = SparkContext(conf=conf)
rdd = sc.parallelize(["yyh","hl","grq","zxj","cby","wfe","mrr","qjy"])
print(rdd.count())
二、输出到文件中
1.saveAsTextFile算子
功能:
将RDD的数据写入文本文件中
支持本地写出、 hdfs等文件系统
语法:
rdd = sc.parallelize([1,2,3,4,5])
rdd.saveAsTextFile("../data/output/test.txt")
2.配置Hadoop相关依赖
调用保存文件的算子,需要配置Hadoop依赖
① 下载Hadoop安装包
http://archive.apache.org/dist/hadoop/common/hadoop-3.0.0/hadoop-3.0.0.tar.gz
② 解压到电脑任意位置
③ 在Python代码中使用os模块配置:
os.environ['HADOOP HOME']='HADOOP解压文件夹路径'
E:\python.learning\hadoop分布式相关\hadoop-3.0.0
④ 下载winutils.exe,并放入Hadoop解压文件夹的bin目录内
https://raw.githubusercontent.com/steveloughran/winutils/master/hadoop-3.0.0/bin/winutils.exe
⑤ 下载hadoop.dll,并放入:C:/Windows/System32 文件夹内
https://raw.githubusercontent.com/steveloughran/winutils/master/hadoop-3.0.0/bin/hadoop.dll
3.代码示例
from pyspark import SparkConf,SparkContext
import os
conf = SparkConf().setMaster("local").setAppName("test_spark")
os.environ['PYSPARK_PYTHON'] = "E:/python.learning/pyt/scripts/python.exe"
sc = SparkContext(conf = conf)
# 准备RDD1
rdd1 = sc.parallelize([1,2,3,4,5])
# 准备RDD2
rdd2 = sc.parallelize([("Hello, 3"),("Spark", 5),("Hi", 7)])
# 准备RDD3
rdd3 = sc.parallelize([[1, 3, 5],[6, 7, 9],[11, 13, 11]])
# 输出到文件中
rdd1.saveAsTextFile("E:\python.learning\hadoop分布式相关\data\output1/rdd1")
rdd2.saveAsTextFile("E:\python.learning\hadoop分布式相关\data\output2/rdd2")
rdd3.saveAsTextFile("E:\python.learning\hadoop分布式相关\data\output3/rdd3")
注:如果输出路径的文件存在,代码将会报错
4.运行结果
创建几个文件取决于Hadoop上的分区数量
解决方式:修改rdd的分区
5.修改rdd分区为1个
方式1
Sparkconf对象设置属性全局并行度为1:
from pyspark import SparkConf, SparkContext
import os
os.environ['PYSPARK_PYTHON'] = "E:/python.learning/pyt/scripts/python.exe"
os.environ['HADOOP_HOME'] = "E:\python.learning\hadoop分布式相关\hadoop-3.0.0"
conf = SparkConf().setMaster("local").setAppName("test_spark")
conf.set("spark.default.parallelize", "1")
sc = SparkContext(conf = conf)
# 准备RDD1
rdd1 = sc.parallelize([1,2,3,4,5])
# 准备RDD2
rdd2 = sc.parallelize([("Hello, 3"),("Spark", 5),("Hi", 7)])
# 准备RDD3
rdd3 = sc.parallelize([[1, 3, 5],[6, 7, 9],[11, 13, 11]])
# 输出到文件中
rdd1.saveAsTextFile("E:\python.learning\hadoop分布式相关\data\output1/rdd1")
rdd2.saveAsTextFile("E:\python.learning\hadoop分布式相关\data\output2/rdd2")
rdd3.saveAsTextFile("E:\python.learning\hadoop分布式相关\data\output3/rdd3")
方式2
创建RDD的时候设置 parallelize方法传入numSlices参数为1:
rdd1 = sc.parallelize([1,2,3,4,5],1)