如何在 Pig 中加入两个数据集并分组后找到平均值

How to find average after joining two datasets and grouping, in Pig

我有两个数据集,包含 4 列(id、姓名、性别、位置)的 EmployeeDetail 和 SalaryDetail(id, salary)。我加入了两个数据集并将它们分组为位置。

EmpDetail = load '/Users/bmohanty6/EmployeeDetails/EmpDetail.txt' as (id:int, name:chararray, gender:chararray, location:chararray);
SalaryDetail = load '/Users/bmohanty6/EmployeeDetails/EmpSalary.txt' as (id:int, salary:float);                                     
JoinedEmpDetail = join EmpDetail by id, SalaryDetail by id;                                                                         
GroupedByLocation = group JoinedEmpDetail by location; 

DUMP GroupedByLocation 给了我预期的正确结果。现在,当我尝试使用下面的行取平均值时,

AverageSalary = foreach GroupedByLocation generate group, AVG(SalaryDetail.salary);

抛出以下错误。

<line 11, column 58> Could not infer the matching function for org.apache.pig.builtin.AVG as multiple or none of them fit. Please use an explicit cast.

我也试过下面的方法。但是得到了同样的错误。

AverageSalary = foreach GroupedByLocation {
  Sum = SUM(SalaryDetail.salary);
  Count = COUNT(SalaryDetail.salary);
  avgSal = Sum/Count;
  generate group as location, avgSal;
  };

这次错误是:

Could not infer the matching function for org.apache.pig.builtin.SUM as multiple or none of them fit. Please use an explicit cast.

谁能告诉我正确的方法。

感谢 Sivasakthi Jayaraman 回答我的问题。

AverageSalary = foreach GroupedByLocation generate group, AVG(JoinedEmpDetail.SalaryDetail::salary);

这给出了每个地点的平均工资。 现在我试图找出每个 location 中每个性别的平均工资。所以我尝试在 GroupedByLocation 变量中按 gender 分组。但面临一些问题。

GroupdByGender = foreach GroupedByLocation { 
genderGrp = group JoinedEmpDetail by JoinedEmpDetail.EmpDetail::gender;
avgSalary = foreach genderGrp generate group, AVG(JoinedEmpDetail.SalaryDetail::salary);
generate group as location, JoinedEmpDetail.EmpDetail::gender, avgSalary;
};

我遇到了这个错误

Syntax error, unexpected symbol at or near 'JoinedEmpDetail'

谁能帮忙。

您不能像这样访问 salary 列,首先您需要映射 JoinedEmpDetail 关系,然后访问 salary 列。

你能试试下面的stmt吗?

AverageSalary = foreach GroupedByLocation generate group, AVG(JoinedEmpDetail.SalaryDetail::salary);