Question

我需要选择那些计划编号为null且不为null的患者，这是我的努力：

select distinct PatientID from Visits where plannumber is null and 
patientid = (
select distinct patientid from Visits where plannumber is not 
null group by PatientID) group by PatientID

返回并返回错误'子查询返回的值超过1'。

有没有更有效的方法来编写此查询？

Answer 1

select distinct V1.PatientID 
  from Visits V1
  join Visits V2 
    on V2.PatientID = V1.PatientID 
   and V1.plannumber is not null 
   and V2.plannumber is null

或

select distinct PatientID 
  from Visits 
 where plannumber is not null
INTERSECT 
select distinct PatientID 
  from Visits 
 where plannumber is null

至于您的查询

patientid in

不是

patientid =

你的独特和分组是多余的

Answer 2

有几种方法，这里有一对......

SELECT
  PatientID
FROM
  Visits
GROUP BY
  PatientID
HAVING
      SUM(CASE WHEN plannumber IS NULL THEN 1 ELSE 0 END) > 0
  AND SUM(CASE WHEN plannumber IS NULL THEN 0 ELSE 1 END) > 0

或者...

SELECT
  *
FROM
  Patient
WHERE
      EXISTS (SELECT * FROM Visits WHERE plannumber IS     NULL AND PatientID = Patient.ID)
  AND EXISTS (SELECT * FROM Visits WHERE plannumber IS NOT NULL AND PatientID = Patient.ID)

写入查询的最有效方法是选择null而不是null

2 个答案: