用Spark做求平均成绩算法
##由于才开始学,此做法为只为结果,不为过程型
实验数据:
math.txt: English.txt:
Ben 98 Ben 89
Bean 99 Bean 98
Harry 89 Harry 78
Sam 79 Sam 87
Tom 80 Tom 80
from pyspark import SparkContext
#定义的函数,用于把两个文件的list合并转换成(Tome,80)形式def TurnDict(dict1):
keys = []
values =[]
for i in range(0,len(dict1)):
if i % 2 != 0:
values.append(int(dict1[i]))
else:
keys.append(dict1[i])
dict2 = zip(keys,values)
return dict2
sc = SparkContext('local','AvGrade')
#创建RDD
rdd1 = sc.textFile("file:///usr/local/spark/mycode/TestPackage/math.txt")
rdd2 = sc.textFile("file:///usr/local/spark/mycode/TestPackage/English.txt")
#对与数据流按" "切割,并且调用map函数转换成(Tom ,1)形式,在接着用keys把key值取出来,再用collect()单独为一个列表pairRDD1 = rdd1.flatMap(lambda line : (line.split(" ")[0],line.split(" ")[1])).map(lambda x : (x,1)).keys().collect()
pairRDD2 = rdd2.flatMap(lambda line : (line.split(" ")[0],line.split(" ")[1])).map(lambda x : (x,1)).keys().collect()
把字典转换成DataFramepairRDD_1 = sc.parallelize(TurnDict(pairRDD1))
pairRDD_2 = sc.parallelize(TurnDict(pairRDD2))
'''
pairRDD = pairRDD_1.join(pairRDD_2)
pairRDD.reduceByKey(lambda x,y : (x+y)).foreach(print)
'''
#把两个RDD里面的值合并为一个列表
pairRDD_1_N = pairRDD_1.collect()
pairRDD_2_N = pairRDD_2.collect()
for i in pairRDD_1_N:
pairRDD_2_N.append(i)
#再转换成RDD
pairRDD = sc.parallelize(pairRDD_2_N)
#调用reducByKey进行计算
result =pairRDD.reduceByKey(lambda x,y : ((x+y)/2))
#打印值
result.foreach(print)