在 case 语句中使用计数函数答案

【问题标题】：Use a count function in a case statment在 case 语句中使用计数函数
【发布时间】：2021-08-26 17:22:54
【问题描述】：

我有下面的代码，我需要在 case 语句的一个条件中输入一个计数函数，然后按下面的 case 语句分组。但这并没有让我对案例陈述进行分组。我基本上想创建一个名为“Both”的第三个存储桶，条件如下

select  
CASE
  WHEN i.CLASSIFICATION in ('CUSTOMER','PARTNER') THEN 'API'
  WHEN i.CLASSIFICATION in ('MOBILE','NDSE') THEN 'APP' 
  WHEN (i.CLASSIFICATION in ('MOBILE','NDSE','CUSTOMER','PARTNER') OR count(distinct UPPER(p.ACCOUNTID)) >=2) THEN 'BOTH' 
  ELSE 'Other' END AS "Bucket", 
count(distinct ACCOUNTID) as No_Accounts,count(distinct envelopeid_hash) as numenvelopes
from ENVELOPE_TABLE e   
JOIN DIMDATE_TABLE b on to_date(e.sentinitial)=b.standarddate
join PAYING_TABLE p
on UPPER(e.ACCOUNTID)=UPPER(p.ACCOUNTID) 
LEFT JOIN IK_TABLE i ON i.IntegratorKeyId = e.InitiatingIntegratorKeyId AND i.SourceKey = e.SourceKey
group by 1

【问题讨论】：

用您正在使用的数据库标记您的问题。

标签： sql if-statement group-by count case

【解决方案1】：

你有第一列的表达式，'select'在group by之后执行，所以你得到一个错误。您需要在group by 中再次重复case when：

select  
CASE
  WHEN i.CLASSIFICATION in ('CUSTOMER','PARTNER') THEN 'API'
  WHEN i.CLASSIFICATION in ('MOBILE','NDSE') THEN 'APP' 
  WHEN (i.CLASSIFICATION in ('MOBILE','NDSE','CUSTOMER','PARTNER') OR count(distinct UPPER(p.ACCOUNTID)) >=2) THEN 'BOTH' 
  ELSE 'Other' END AS "Bucket", 
count(distinct ACCOUNTID) as No_Accounts,count(distinct envelopeid_hash) as numenvelopes
from ENVELOPE_TABLE e   
JOIN DIMDATE_TABLE b on to_date(e.sentinitial)=b.standarddate
join PAYING_TABLE p
on UPPER(e.ACCOUNTID)=UPPER(p.ACCOUNTID) 
LEFT JOIN IK_TABLE i ON i.IntegratorKeyId = e.InitiatingIntegratorKeyId AND i.SourceKey = e.SourceKey
group by CASE
  WHEN i.CLASSIFICATION in ('CUSTOMER','PARTNER') THEN 'API'
  WHEN i.CLASSIFICATION in ('MOBILE','NDSE') THEN 'APP' 
  WHEN (i.CLASSIFICATION in ('MOBILE','NDSE','CUSTOMER','PARTNER') OR count(distinct UPPER(p.ACCOUNTID)) >=2) THEN 'BOTH' 
  ELSE 'Other' END AS "Bucket"

【讨论】：