Sql server SQL Server计数查询将结果与总数进行比较

我试图返回被完全拒绝的不同发票的数量。一张发票可以有多个行项目。如果发票中的行项目的原因为“无”，而另一行项目的原因为“被拒绝”，则该发票不是完全被拒绝的，不应计算在内

我想要的结果如下

Client     | No. of Invoices Fully Rejected
-------------------------------------------
Schogotten | 1
Maiken     | 2

从逻辑上讲，我认为我返回的发票数量中，行项目的数量等于每张发票拒绝原因的发票数量

这是你的电话号码

MS SQL Server 2008架构设置：

CREATE TABLE Tmp (
  id int NOT NULL,
  client varchar (24) NOT NULL,
  invoice varchar(5) NOT NULL,
  statusid int NOT NULL,
  reason varchar (24)
)


INSERT INTO Tmp (id, client, invoice, statusid, reason) 
VALUES (12345, 'Schogotten', 'ABC12', -99, 'Rejected')
, (12346, 'Schogotten', 'ABC12', -99, 'None')
, (12347, 'Schogotten', 'ABC12', 10001, 'ReadyToPay')
, (12348, 'Schogotten', 'ABC11', -99, 'Rejected')
, (12349, 'Schogotten', 'ABC10', -99, 'None')
, (12350, 'Maiken', 'ABC13', -99, 'Rejected')
, (12351, 'Maiken', 'ABC14', 9006, 'ReadyToPay')
, (12351, 'Maiken', 'ABC14', -99, 'None')
, (12352, 'Maiken', 'ABC15', -99, 'Rejected');

SELECT Client
, COUNT(Invoice)
FROM Tmp
WHERE Reason = 'Rejected'
GROUP BY Client

|     CLIENT | COLUMN_1 |
|------------|----------|
|     Maiken |        2 |
| Schogotten |        2 |

查询1：

CREATE TABLE Tmp (
  id int NOT NULL,
  client varchar (24) NOT NULL,
  invoice varchar(5) NOT NULL,
  statusid int NOT NULL,
  reason varchar (24)
)


INSERT INTO Tmp (id, client, invoice, statusid, reason) 
VALUES (12345, 'Schogotten', 'ABC12', -99, 'Rejected')
, (12346, 'Schogotten', 'ABC12', -99, 'None')
, (12347, 'Schogotten', 'ABC12', 10001, 'ReadyToPay')
, (12348, 'Schogotten', 'ABC11', -99, 'Rejected')
, (12349, 'Schogotten', 'ABC10', -99, 'None')
, (12350, 'Maiken', 'ABC13', -99, 'Rejected')
, (12351, 'Maiken', 'ABC14', 9006, 'ReadyToPay')
, (12351, 'Maiken', 'ABC14', -99, 'None')
, (12352, 'Maiken', 'ABC15', -99, 'Rejected');

SELECT Client
, COUNT(Invoice)
FROM Tmp
WHERE Reason = 'Rejected'
GROUP BY Client

|     CLIENT | COLUMN_1 |
|------------|----------|
|     Maiken |        2 |
| Schogotten |        2 |

：

CREATE TABLE Tmp (
  id int NOT NULL,
  client varchar (24) NOT NULL,
  invoice varchar(5) NOT NULL,
  statusid int NOT NULL,
  reason varchar (24)
)


INSERT INTO Tmp (id, client, invoice, statusid, reason) 
VALUES (12345, 'Schogotten', 'ABC12', -99, 'Rejected')
, (12346, 'Schogotten', 'ABC12', -99, 'None')
, (12347, 'Schogotten', 'ABC12', 10001, 'ReadyToPay')
, (12348, 'Schogotten', 'ABC11', -99, 'Rejected')
, (12349, 'Schogotten', 'ABC10', -99, 'None')
, (12350, 'Maiken', 'ABC13', -99, 'Rejected')
, (12351, 'Maiken', 'ABC14', 9006, 'ReadyToPay')
, (12351, 'Maiken', 'ABC14', -99, 'None')
, (12352, 'Maiken', 'ABC15', -99, 'Rejected');

SELECT Client
, COUNT(Invoice)
FROM Tmp
WHERE Reason = 'Rejected'
GROUP BY Client

|     CLIENT | COLUMN_1 |
|------------|----------|
|     Maiken |        2 |
| Schogotten |        2 |

---

我将使用子查询查找相关发票，然后使用外部查询按客户分组：

select client,COUNT(*) from (
  select client,invoice
  from Tmp
  group by client,invoice
  having MIN(reason) = MAX(reason) and
  MIN(reason) = 'Rejected'
) t
group by client

---

希望逻辑很容易理解-子查询只查找所有行具有相同值（

MIN（reason）=MAX（reason）

）的发票，其中该原因被拒绝，然后我们只对其进行计数。

您知道SQL fiddle在“运行SQL”上有一个“Markdown-uptut”选项吗按钮，你可以简单地粘贴到上述问题中？@Tanner我现在是。谢谢你引起我的注意。谢谢，这很有道理，是解决我问题的办法。