带大小写和求和的 Spark SQL

Question

我有一个配置单元表，其中有列（id、dept、salary）我正在使用 spark SQL 在该表上应用一些转换。

即如果部门是 HR 那么salary=salary+100，如果部门是 IT 那么salary=salary+0 并且基于部门取薪水的总和。光纤通道

输入数据

+----+-----------+--------+
| id |   dept    | salary |
+----+-----------+--------+
|  1 | HR        |    100 |
|  2 | operation |    200 |
|  3 | tech      |    300 |
|  4 | IT        |    400 |
|  1 | HR        |    500 |
+----+-----------+--------+

预期输出

+----+-----------+--------+
| id |   dept    | salary |
+----+-----------+--------+
|  1 | HR        |    800 |
|  2 | operation |    300 |
|  3 | tech      |    300 |
|  4 | IT        |    400 |
+----+-----------+--------+

我已经编写了下面的代码，但它在 spark sql 中不起作用。

spark.sql("select CASE WHEN dept = 'HR' THEN 'sum(salary+100)',when  dept = 'IT' THEN 'sum(salary+0)' ELSE 'salary' END AS salary from emp group by dept").show

但输入不匹配：问题。请帮助我应该如何实现这一点。

Answer 1

预期输出：spark sql 查询是

spark.sql("Select id,dept, sum(salary) + sum(case dept when 'HR' then 100 when 'IT' then 0 when 'operation' then 100 else 0 end) as salary  From employee1 group by id,dept order by id asc").show()

我希望这会有所帮助。

Answer 2

Select Dept, sum(salary + case Dept when 'HR' then 100 when 'IT' then 0 else 0 end)
From emp
Group by Dept

或者

Select Dept, sum(salary) + sum(case Dept when 'HR' then 100 when 'IT' then 0 else 0 end)
From emp
Group by Dept