Sql 在不使用join的情况下合并3个表，但需要分组

我有3个不同的表格，如下所示：

| DataEntry:                   |  | OOW:               |  | ContractExpired:   |
| +----+-----------------------|  | +----+-------------|  | +----+-------------|
| Country | valid  | invalid   |  | Country | Warranty |  | Country | Expired  |
| AU      | 1      | 2         |  | AU      | 1        |  | AU      | 1        |
| CN      | 22     | 3         |  | CN      | 1        |  | CN      | 1        |
| JP      | 1      | 1         |  | +----+-------------|  | +----+-------------|
| KR      | 1      | 1         |
| SG      | 1      | 1         |
| +----+-----------------------|

我写的查询如下：

SELECT Country,
(SELECT SUM(CASE WHEN CallType = 'valid' THEN 1 ELSE Null END) AS [valid] FROM DataEntry AS DataEntry_1) AS [valid],
(SELECT SUM(CASE WHEN CallType = 'Invalid' THEN 1 ELSE Null END) AS [Invalid] FROM DataEntry AS DataEntry_2) AS [Invalid],
(SELECT COUNT(CallType) AS [Warranty] FROM OOW) AS [Warranty],
(SELECT COUNT(CallType) AS [Expired] FROM ContractExpired) AS [Expired]

返回的结果如下所示：

| valid | invalid | warranty | Expired |
| 26    | 8       | 2        | 2       |

我想通过分组到国家对其进行一些更改： 样品

我尝试按分组，但查询返回错误结果：

SELECT Country,
(SELECT SUM(CASE WHEN CallType = 'valid' THEN 1 ELSE Null END) AS [valid] FROM DataEntry AS DataEntry_1) AS [valid],
(SELECT SUM(CASE WHEN CallType = 'Invalid' THEN 1 ELSE Null END) AS [Invalid] FROM DataEntry AS DataEntry_2) AS [Invalid],
(SELECT COUNT(CallType) AS [Warranty] FROM OOW) AS [Warranty],
(SELECT COUNT(CallType) AS [Expired] FROM ContractExpired) AS [Expired]
FROM DataEntry AS DataEntry_3
Group by Country

结果:

| Country | valid | invalid | warranty | Expired |
| AU      | 26    | 8       | 2        | 2       |
| CN      | 26    | 8       | 2        | 2       |
| JP      | 26    | 8       | 2        | 2       |
| KR      | 26    | 8       | 2        | 2       |
| SG      | 26    | 8       | 2        | 2       |

---

请提供一些指导来解决这个问题。

您的查询不应该是下面这样的吗。这可能不是确切的查询，但会让您知道如何使其工作

SELECT de.Country,
SUM(CASE WHEN de.CallType = 'valid' THEN 1 ELSE Null END) AS [valid],
SUM(CASE WHEN de.CallType = 'Invalid' THEN 1 ELSE Null END) AS [Invalid],
FROM DataEntry de
join
(
select COUNT(CallType) AS [Warranty] 
FROM OOW
group by country
) tab on de.country = tab.country
join
(
SELECT COUNT(CallType) AS [Expired] 
FROM ContractExpired
group by country
) tab1 on de.country = tab1.country
Group by de.Country

---

您需要在子查询中执行group by，然后使用外部联接按国家/地区代码将聚合数据联接在一起，因为并非所有国家/地区的OOW和ContractExpired表中都有条目。请注意，这将仅显示数据条目表中至少有一条有效或无效记录的国家/地区

SELECT [countryDataEntry].Country, [countryDataEntry].[valid], [countryDataEntry].[Invalid], Warranty, Expired
from ((SELECT Country
            ,SUM(CASE WHEN CallType = 'valid' THEN 1 ELSE Null END) AS [valid]
            ,SUM(CASE WHEN CallType = 'Invalid' THEN 1 ELSE Null END) as [Invalid] 
       FROM DataEntry AS DataEntry_1 GROUP BY Country
      ) AS [countryDataEntry]
    LEFT OUTER JOIN (SELECT Country, COUNT(CallType) AS [Warranty] FROM OOW GROUP BY Country) AS [Warranty] ON [countryDataEntry].Country = [Warranty].Country)
    LEFT OUTER JOIN (SELECT Country, COUNT(CallType) AS [Expired] FROM ContractExpired GROUP BY Country) AS [Expired] ON [countryDataEntry].Country = [Expired].Country

试试这个：

SELECT * FROM DataEntry DE
LEFT OUTER join OOW ON
OOW.Country = DE.Country
LEFT OUTER JOIN ContractExpired CE ON
CE.Country = DE.Country

如果您想查看以下SQLFIDLE的数据示例，则其仅在3天内处于活动状态：

---

如果您真的想在没有联接的情况下执行此操作，您可以，而且非常简单。您必须使用子选择，并使用WHERE子句对其进行筛选：

SELECT de.Country
  SUM(CASE WHEN de.CallType = 'valid'   THEN 1 ELSE Null END) AS [valid],
  SUM(CASE WHEN de.CallType = 'Invalid' THEN 1 ELSE Null END) AS [Invalid],
  (SELECT COUNT(CallType)
    FROM OOW o
    WHERE o.Country=de.Country) AS [Warranty],
  (SELECT COUNT(CallType)
    FROM ContractExpired ce
    WHERE ce.Country=de.Country) AS [Expired]
FROM DataEntry de
GROUP BY de.Country

有一个问题

此SQL假定您具有与SQL匹配的以下表结构，但不是您提供给我们的结构：

DataEntry(
  Country  CHAR,
  Calltype CHAR);

OOW(
  Country  CHAR,
  Calltype CHAR);

CREATE TABLE ContractExpired(
  Country  CHAR,
  Calltype CHAR);

---

确保在所有三个表中都有关于国家的索引。但对于这里发布的任何答案，无论是连接还是子选择，这都是一个很好的建议。

这可能是因为您使用的是求和和和计数吗？合并查询表怎么样？您应该先用group by创建四个子查询，然后将它们连接在一起，而不是在select子句中使用子查询。@Jaugar Chang:我试过了，但这样做，数据会重复。您应该在join语句和结果中，然后我们可以找出问题所在。实际上，您的SQL与您的数据库结构不匹配。我想知道数据输入表。每个国家/地区是否有一行，其中包含所示结构中的有效和无效调用类型之和，或者是否有多行，其中的calltype列告诉您调用是否有效，如您使用的SQL中所示？要添加到Rahul的答案中，请确保您使用的是外部联接，否则您将不会获得JP、KR或SG的任何记录。谢谢！这就是答案。我使用的是交叉连接。未调用特定字段。难怪它不起作用，重复数据。它也起作用。我没有调用函数->table1.country=table2.country。这就是为什么数据在假设为空时不显示null。谢谢

DataEntry(
  Country  CHAR,
  Calltype CHAR);

OOW(
  Country  CHAR,
  Calltype CHAR);

CREATE TABLE ContractExpired(
  Country  CHAR,
  Calltype CHAR);