Question

我想选择在过去3个月内签署了大部分合同的员工：表之间的连接是：contracts.responsible references employees（id）

select b.name, b.id, count(*) nr_contracts_last_3_mths
    from contracts a join employees b
    on a.responsible = b.id and   
    data between add_months(trunc(sysdate, 'MM'), -1) and trunc(sysdate)
    group by b.name, b.id
    having count(*) = 
      (select max(count(responsible)) 
            from contracts, employees
              where contracts.responsible = employees.id
              group by employees.name
        );

结果仅包含空的选定列（无数据）。

预期结果应该响应的名称，他的ID 和值 5 ，因为bellow查询返回5;

select max(count(*)) nr_contracts_last_3_mths from contracts where data 
       between add_months(trunc(sysdate, 'MM'), -3) and trunc(sysdate)
       group by data;

Answer 1

您的查询不起作用，因为having子句中的子查询在时间范围内没有条件，因此它返回的计数高于“top”查询中的任何行（如，据推测，员工在过去三个月之前也签订了合同。）

无论如何，更简单的方法是将窗口rank功能与“顶级”查询结合使用：

SELECT name, id, nr_contracts_last_3_mths
FROM   (SELECT   b.name AS name, 
                 b.id AS id,
                 COUNT(*) AS nr_contracts_last_3_mths
                 RANK() OVER (ORDER BY COUNT(*) DESC) AS rk
        FROM     contracts a
        JOIN     employees b ON a.responsible = b.id AND
                                data BETWEEN 
                                     add_months(trunc(sysdate, 'MM'), -1) AND 
                                     TRUNC(sysdate)
        GROUP BY b.name, b.id)
WHERE  rk = 1

Answer 2

select b.name, b.id, count(*) nr_contracts_last_3_mths
    from contracts a join employees b
    on a.responsible = b.id and   
    data between add_months(trunc(sysdate, 'MM'), -3) and trunc(sysdate)
    group by b.name, b.id
    having count(*) = 
      (select max(count(*)) 
        from contracts a join employees b
           on a.responsible = b.id and 
           data between add_months(trunc(sysdate, 'MM'), -3) and trunc(sysdate)
           group by b.name, b.id
      );

最大（数）oracle

2 个答案: