練習:
1、將下列JSON格式數據復制到Linux系統中,並保存命名為employee.json。
| { "id":1 , "name":" Ella" , "age":36 } { "id":2, "name":"Bob","age":29 } { "id":3 , "name":"Jack","age":29 } { "id":4 , "name":"Jim","age":28 } { "id":4 , "name":"Jim","age":28 } { "id":5 , "name":"Damon" } { "id":5 , "name":"Damon" } |
為employee.json創建DataFrame,並寫出Python語句完成下列操作:
>>spark=SparkSession.builder.getOrCreate()
>>df=spark.read.json('file:///usr/local/spark/mycode/dafaframe/employee.json')
1.查詢所有數據;
>>df.show()

2.查詢所有數據,並去除重復的數據;
>>df.distinct().show()

3.查詢所有數據,打印時去除id字段;
>>df.select(df.name,df.age).show()

4.篩選出age>30的記錄;
>>df.filter(df.age>30).show()

5.將數據按age分組;
>>df.groupby('age').count().show()

6.將數據按name升序排列;
>>df.sort(df.name.asc()).show()

7.取出前3行數據;
>>df.show(3)

8.查詢所有記錄的name列,並為其取別名為username;
>>df.select(df.name.alias('username')).show()

9.查詢年齡age的平均值;
>>df.groupby.avg('age').collect()[0].adDict()['avg(age)']
30.0
10.查詢年齡age的最小值。
>>df.groupby.min('age').collect()[0].adDict()['min(age)']
28
2.編程實現將RDD轉換為DataFrame
源文件內容如下(包含id,name,age):
| 1,Ella,36 2,Bob,29 3,Jack,29 |
請將數據復制保存到Linux系統中,命名為employee.txt,實現從RDD轉換得到DataFrame,並按“id:1,name:Ella,age:36”的格式打印出DataFrame的所有數據。請寫出程序代碼。
from pyspark.sql import SparkSession, Row
spark = SparkSession.builder.appName('employee').getOrCreate()
sc = spark.sparkContext
lines = sc.textFile('file:///usr/local/spark/mycode/data/employee.txt')
result1 = lines.filter(lambda line: (len(line.strip()) > 0))
result2 = result1.map(lambda x: x.split(','))
#將RDD轉換成DataFrame
item = result2.map(lambda x: Row(id=x[0], name=x[1], age=x[2]))
df = spark.createDataFrame(item)
df.show()

3.統計chines_year文件每年各類節目的數量,打印(節目名稱、數量、年份)。要求首先按照節目名稱升序排序,節目名稱相同時其次按照年份升序排序。采用Spark RDD和Spark SQL兩種方式。分別寫出代碼並截圖。
sortTypeRDD.py
from pyspark import SparkConf, SparkContext
from operator import gt
class SecondSortKey():
def __init__(self, k):
self.year = k[0]
self.name = k[1]
def __gt__(self, other):
if other.name == self.name:
return gt(self.year, other.year)
else:
return gt(self.name, other.name)
def main():
conf = SparkConf().setAppName('spark_sort').setMaster('local')
sc = SparkContext(conf=conf)
data = sc.textFile('file:///usr/local/spark/mycode/data/chinese_year.txt')
rdd = data .map(lambda x: (x.split("\t")[1], x.split("\t")[0]))\
.map(lambda x: (x, 1))\
.reduceByKey(lambda a, b: a+b)\
.map(lambda x: (SecondSortKey(x[0]), x[0][0]+','+x[0][1]+','+str(x[1])))\
.sortByKey(True)\
.map(lambda x: x[1])
rdd.foreach(print)
if __name__=='__main__':
main()
SortTypeSql.py
from pyspark.sql import SparkSession, Row
#讀取text文件
spark = SparkSession.builder.appName('topNSQL').getOrCreate()
sc = spark.sparkContext
lines = sc.textFile("file:///usr/local/spark/mycode/data/chinese_year.txt")
result1 = lines.filter(lambda line: (len(line.strip()) > 0) and (len(line.split('\t'))==4))
result2 = result1.map(lambda x: x.split('\t'))
#將RDD轉換成DataFrame
item = result2.map(lambda x: Row(year=x[0], type=x[1], program=x[2], performers=x[3]))
df = spark.createDataFrame(item)
df.createOrReplaceTempView('items')
df1 = spark.sql('select type,year,count(*) from items group by type ,year order by type ,year ')
df1.show()

