sparkSQL面试题

一、查询所有数学课程成绩大于语文课程成绩的学生学号

数据

1,yuwen,43
1,shuxue,55
2,yuwen,77
2,shuxue,88
3,yuwen,98
3,shuxue,65
3,yingyu,88

基本步骤：

• 进行行转列
• 比较语文与数学的成绩

SQL代码： 

with t1 as(
        SELECT id,sum(if(name = 'yuwen',score,0)) chinese,sum(if(name = 'shuxue',score,0)) math,sum(if(name = 'yingyu',score,0)) english FROM zuoye1 group by id
    )
    select id from t1 where math>chinese

 sparkSQL代码

import os
import re

from pyspark.sql import SparkSession

"""
------------------------------------------
  Description : TODO：
  SourceFile : _04-zuoye1
  Author  : zxx
  Date  : 2024/11/4
-------------------------------------------
"""
# 查询所有数学课程成绩大于语文课程成绩的学生学号
if __name__ == '__main__':
    os.environ['JAVA_HOME'] = 'D:/bigdata/03-java/java-8/jdk'
    # 配置Hadoop的路径，就是前面解压的那个路径
    os.environ['HADOOP_HOME'] = 'D:/bigdata/04-Hadoop/hadoop/hadoop-3.3.1/hadoop-3.3.1'
    # 配置base环境Python解析器的路径
    os.environ['PYSPARK_PYTHON'] = 'D:/bigdata/22-spark/Miniconda3/python.exe'  # 配置base环境Python解析器的路径
    os.environ['PYSPARK_DRIVER_PYTHON'] = 'D:/bigdata/22-spark/Miniconda3/python.exe'
    spark = SparkSession.builder.master("local[2]").appName("第一题").config(
        "spark.sql.shuffle.partitions", 2).getOrCreate()
    df = (spark.sparkContext.textFile("../../datas/zuoye2/zuoye1.txt")
          .map(lambda line: (re.split(",",line)[0],re.split(",",line)[1],re.split(",",line)[2])).toDF(["id","name","score"]))
    df.createOrReplaceTempView("zuoye1")
    spark.sql("""
    with t1 as(
        SELECT id,sum(if(name = 'yuwen',score,0)) chinese,sum(if(name = 'shuxue',score,0)) math,sum(if(name = 'yingyu',score,0)) english FROM zuoye1 group by id
    )
    select id from t1 where math>english
    """).show()
    # 关闭
    spark.stop()

运行结果：

二、查询每种爱好中年龄最大的人，如果有相同的年龄，并列显示。

数据

id;name;age;favor
1;huangbo;33;a,b,c,d,e
2;xuzheng;44;b,c
3;wangbaoqiang;33;c,d,e
4;fanbingbing;32;a,b,d

基本步骤： 

• 把favor列用炸裂函数explode炸开
• 用排名函数进行排序，根据题意相同年龄要并列显示，故选用rank()
• 用where筛选出排名第一的人

SQL代码： 

with t1 as(
   select name,age,aihao from zuoye2 lateral view explode(split(favor,',')) t1 as aihao)
   ,t2 as(
   select aihao,name,rank() over(partition by aihao order by age) r1 from t1
   )
   select * from t2 where r1 = 1 order by aihao

sparkSQL代码：

import os
import re

from pyspark.sql import SparkSession

"""
------------------------------------------
  Description : TODO：
  SourceFile : _04-zuoye1
  Author  : zxx
  Date  : 2024/11/4
-------------------------------------------
"""
# 查询每种爱好中年龄最大的人，如果有相同的年龄，并列显示。
if __name__ == '__main__':
    os.environ['JAVA_HOME'] = 'D:/bigdata/03-java/java-8/jdk'
    # 配置Hadoop的路径，就是前面解压的那个路径
    os.environ['HADOOP_HOME'] = 'D:/bigdata/04-Hadoop/hadoop/hadoop-3.3.1/hadoop-3.3.1'
    # 配置base环境Python解析器的路径
    os.environ['PYSPARK_PYTHON'] = 'D:/bigdata/22-spark/Miniconda3/python.exe'  # 配置base环境Python解析器的路径
    os.environ['PYSPARK_DRIVER_PYTHON'] = 'D:/bigdata/22-spark/Miniconda3/python.exe'
    spark = SparkSession.builder.master("local[2]").appName("第一次构建SparkSession").config(
        "spark.sql.shuffle.partitions", 2).getOrCreate()
    df = (spark.sparkContext.textFile("../../datas/zuoye2/zuoye2.txt").filter(lambda x:re.split(";",x)[0] !="id")
          .map(lambda line: (re.split(";",line)[0],re.split(";",line)[1],re.split(";",line)[2],
                             re.split(";",line)[3]
                             )
               ).toDF(["id","name","age","favor"]))
    df.createOrReplaceTempView("zuoye2")
    spark.sql("""
    with t1 as(
   select name,age,aihao from zuoye2 lateral view explode(split(favor,',')) t1 as aihao)
   ,t2 as(
   select aihao,name,rank() over(partition by aihao order by age) r1 from t1
   )
   select * from t2 where r1 = 1 order by aihao
    """).show()
    # 关闭
    spark.stop()

运行结果：