05 RDD编程
一、词频统计：
1.读文本文件生成RDD lines
lines=sc.textFile("file:///usr/local/spark/mycode/rdd/word.txt")
lines.foreach(print)
2.将一行一行的文本分割成单词 words flatmap()
words=lines.flatMap(lambda line:line.split())
words.foreach(print)
3.全部转换为小写 lower()
wordsxx=lines.map(lambda word:word.lower())
wordsxx.foreach(print)
4.去掉长度小于3的单词 filter()
word=words.filter(lambda words:len(words)>2)
word.foreach(print)
5.去掉停用词
lines=textFile("file:///usr/local/spark/mycode/rdd/word.txt")
with open("/usr/lcaol/spark/mycode/rdd/stopwords.txt") as f:
stops=f.read().split()
lines.flatMap(lambda line:line.split()).filter(lambda word:word not in stops).collect()
6.转换成键值对 map()
words.map(lambda word:(word,1)).collect()
7.统计词频 reduceByKey()
words.map(lambda word:(word,1)).reduceByKey(lambda a,b:a+b).collect()
8.按字母顺序排序 sortBy(f)
words.map(lambda word:(word,1)).reduceByKey(lambda a,b:a+b).sortBy(lambda word:word[0]).collect()
9.按词频排序 sortByKey()
words.map(lambda word:(word,1)).reduceByKey(lambda a,b:a+b).sortByKey().collect()
10.结果文件保存 saveAsTextFile(out_url)
lines.saveAsTextFile('file:///usr/local/spark/mycode/words0.txt')
二、学生课程分数案例
lines=sc.textFile("file:///usr/local/spark/mycode/rdd/xs.txt")
- 总共有多少学生？map(), distinct(), count()
lines.map(lambda line:line.split(',')[0]).distinct().count()
- 开设了多少门课程？
lines.map(lambda line:line.split(',')[1]).distinct().count()
- 每个学生选修了多少门课？map(), countByKey()
lines.map(lambda line:line.split(',')).map(lambda line:(ine[0],line[1])).countByKey()
- 每门课程有多少个学生选？map(), countByValue()
lines.map(lambda line:line.split(',')).map(lambda line:line[1]).countByValue()
- Tom选修了几门课？每门课多少分？filter(), map() RDD
lines.file(lambda line:"Tom" in line).map(lambda line:line.split(','))
- Tom选修了几门课？每门课多少分？map(),lookup() list
lines.map(lambda line:line.split(',')).map(lambda line:(line[0],(line[1],line[2]))).lookup('Tom')
- Tom的成绩按分数大小排序。filter(), map(), sortBy()
lines.filter(lambda line:'Tom' in line).map(lambda line:l;ine.split(',')).sortBy(lambda line:line[2],False).collect()
- Tom的平均分。map(),lookup(),mean()
import numpy as np
name=lines.map(lambda line:line.split(',')).map(lambda line:(line[0],(line[1],line[2]))).lookup('Tom')
np.mean([int(x) for x in name])
- 生成（课程，分数）RDD，观察keys(),values()
lines=sc.textFile('file:///usr/local/spark/mycode/rdd/xs.txt')
km=lines.map(lambda line:line.split()).map(lambda line:(line[1],line[2]))
km.take(10)
km.keys().take(10)
km.values().take(10)
map(lambda line:line.split())
map(lambda line:line.split())
- 每个分数+5分。mapValues(func)
km5=km.mapValues(lambda km:int(km)+5)
km5.take(10)
- 求每门课的选修人数及所有人的总分。combineByKey()
kmrz=km.combineByKey(lambda km:(int(km),1), lambda km,fs:(km[0]+int(fs),km[1]+1), lambda km1,km2:(km1[0]+km2[0],km1[1]+km2[1]))
kmrz.take(10)
- 求每门课的选修人数及平均分，精确到2位小数。map()，round()
kmrz.map(lambda km:(km[0],km[1][1],round(km[1][0]/km[1][1],2))).collect()
- 求每门课的选修人数及平均分。用reduceByKey()实现，并比较与combineByKey()的异同。
km.map(lambda km:(km[0],(int(km[1]),1))).reduceByKey(lambda a,b:(a[0]+b[0],a[1]+b[1])).map(lambda km:(km[0],km[1][1],round(km[1][0]/km[1][1],2))).collect()
