Question

美好时光。

说，有一个表格包含有关考试结果的数据：

                EXAM_RESULTS
| ID | EXAM_TYPE | PERSON_ID | EXAM_RESULT |

学生可以尝试多次通过具体考试，因此一个学生可以在上表中有四行：三个用于失败结果，一个用于成功。我需要选择尚未通过具体考试类型的所有学生。

请建议，如何操作或在何处阅读有关此类技巧的内容。

修改：样本记录：

| 1 | SDA | 111 | FAIL |
| 2 | SDA | 111 | FAIL |
| 3 | SDA | 111 | PASSED |
| 4 | SDA | 222 | FAIL |
| 4 | SDA | 222 | FAIL |

根据任务，查询必须只选择222人因为他没有通过SDA考试（111人最终通过）

Answer 1

SELECT  Person_ID
FROM    Exam_Result
WHERE   Exam_type = 'type_here'
GROUP   BY Person_ID
HAVING  COUNT(CASE WHEN EXAM_RESULT = 'PASSED' THEN 1 END) = 0

Answer 2

/* Students that not have passed the exam: either haven't taken it or failed all their's attempts */
/* All the students */
select distinct Person_id
  from Exam_Results

except /* this keyword is server dependent! MINUS for Oracle*/

/* Students that have passed */
select Person_Id
  from Exam_Results
 where (Exam_Result = 'passed') and
       (Exam_Type = 'my exam type')

Answer 3

只是为了让您先行一步，这就是您可以做的事情：

SELECT PersonId,
       COUNT(CASE WHEN ExamResult = 'PASS' THEN 1 END) CNT
FROM ExamResults
WHERE ExamType = 'CONCRETE'
GROUP BY PersonId
HAVING COUNT(CASE WHEN ExamResult = 'PASS' THEN 1 END) > 0

SQL FIDDLE DEMO

SQL：disjoint选择一个表

3 个答案: