评估同一行的多个条件答案

【问题标题】：Evaluate Multiple conditions for same row评估同一行的多个条件
【发布时间】：2020-11-23 10:06:24
【问题描述】：

我必须比较 2 个不同的来源并找出所有 IDs 的所有不匹配项

Source_excel表

+-----+-------------+------+----------+
| id  | name        | City | flag     |
+-----+-------------+------+----------+
| 101 | Plate       | NY   | Ready    |
| 102 | Back washer | NY   | Sold     |
| 103 | Ring        | MC   | Planning |
| 104 | Glass       | NMC  | Ready    |
| 107 | Cover       | PR   | Ready    |
+-----+-------------+------+----------+

Source_dw表

+-----+----------+------+----------+
| id  | name     | City | flag     |
+-----+----------+------+----------+
| 101 | Plate    | NY   | Planning |
| 102 | Nut      | TN   | Expired  |
| 103 | Ring     | MC   | Planning |
| 104 | Top Wire | NY   | Ready    |
| 105 | Bolt     | MC   | Expired  |
+-----+----------+------+----------+

预期结果

+-----+-------------+----------+------------+----------+------------+---------+------------------+
| ID  | excel_name  | dw_name  | excel_flag | dw_flag  | excel_city | dw_city | RESULT           |
+-----+-------------+----------+------------+----------+------------+---------+------------------+
| 101 | Plate       | Plate    | Ready      | Planning | NY         | NY      | FLAG_MISMATCH    |
| 102 | Back washer | Nut      | Sold       | Expired  | NY         | TN      | NAME_MISMATCH    |
| 102 | Back washer | Nut      | Sold       | Expired  | NY         | TN      | FLAG_MISMATCH    |
| 102 | Back washer | Nut      | Sold       | Expired  | NY         | TN      | CITY_MISMATCH    |
| 103 | Ring        | Ring     | Planning   | Planning | MC         | MC      | ALL_MATCH        |
| 104 | Glass       | Top Wire | Ready      | Ready    | NMC        | NY      | NAME_MISMATCH    |
| 104 | Glass       | Top Wire | Ready      | Ready    | NMC        | NY      | CITY_MISMATCH    |
| 107 | Cover       |          | Ready      |          | PR         |         | MISSING IN DW    |
| 105 |             | Bolt     |            | Expired  |            | MC      | MISSING IN EXCEL |
+-----+-------------+----------+------------+----------+------------+---------+------------------+

我已经尝试了下面的查询，但它只给出了一个不匹配的结果。

select ISNULL(EXCEL.ID,DW.ID) ID, 
       excel.name as excel_name,dw.name as dw_name,
       excel.flag as excel_flag,dw.flag as dw_flag,
       excel.city as excel_city,dw.city as dw_city,
       RESULT = CASE WHEN excel.ID IS NULL THEN 'MISSING IN EXCEL'
                     WHEN dw.ID IS NULL THEN 'MISSING IN DW'
                     WHEN excel.NAME<>dw.NAME THEN 'NAME_MISMATCH'
                     WHEN excel.CITY<>dw.CITY THEN 'CITY_MISMATCH'
                     WHEN excel.FLAG <> dw.FLAG THEN 'FLAG_MISMATCH'
                ELSE 'ALL_MATCH' END
from source_excel excel
FULL OUTER JOIN source_dw dw ON excel.id=dw.id

实际输出

+-----+-------------+----------+------------+----------+------------+---------+------------------+
| ID  | excel_name  | dw_name  | excel_flag | dw_flag  | excel_city | dw_city | RESULT           |
+-----+-------------+----------+------------+----------+------------+---------+------------------+
| 101 | Plate       | Plate    | Ready      | Planning | NY         | NY      | FLAG_MISMATCH    |
| 102 | Back washer | Nut      | Sold       | Expired  | NY         | TN      | NAME_MISMATCH    |
| 103 | Ring        | Ring     | Planning   | Planning | MC         | MC      | ALL_MATCH        |
| 104 | Glass       | Top Wire | Ready      | Ready    | NMC        | NY      | NAME_MISMATCH    |
| 107 | Cover       |          | Ready      |          | PR         |         | MISSING IN DW    |
| 105 |             | Bolt     |            | Expired  |            | MC      | MISSING IN EXCEL |
+-----+-------------+----------+------------+----------+------------+---------+------------------+

我知道case 表达式只会检查第一个满足条件。有没有其他方法可以检查所有情况？

【问题讨论】：

您当前的输出在我看来是正确的，至少这是我在这里所期望的。为什么 id 102 和 104 应该重复/三次？
对于每个不匹配，我想要单独的行@TimBiegeleisen。就像 FLAG_MISMATCH 和 CITY_MISMATCH 在单独的行中一样。在这里我只得到一个不匹配

标签： sql sql-server case unpivot full-outer-join

【解决方案1】：

我会将不匹配的地方连接成一行，将原因连接在一起：

select COALESCE(EXCEL.ID, DW.ID) as ID, 
       excel.name as excel_name,dw.name as dw_name,
       excel.flag as excel_flag,dw.flag as dw_flag,
       excel.city as excel_city,dw.city as dw_city,
       (CASE WHEN excel.ID IS NULL
             THEN 'MISSING IN EXCEL'
             WHEN dw.ID IS NULL
             THEN 'MISSING IN DW'
             WHEN excel.NAME = dw.NAME AND excel.CITY = dw.CITY AND excel.FLAG = dw.FLAG
             THEN 'ALL MATCH'
             ELSE CONCAT(CASE WHEN excel.NAME <> dw.NAME THEN 'NAME_MISMATCH; ' END,
                         CASE WHEN excel.CITY <> dw.CITY THEN 'CITY_MISMATCH; ' END,
                         CASE WHEN excel.FLAG <> dw.FLAG THEN 'FLAG_MISMATCH;' END
                        )
        END)
from source_excel excel FULL OUTER JOIN
     source_dw dw
     ON excel.id = dw.id;

【讨论】：

【解决方案2】：

如果我没听错的话，你希望每个不匹配出现一行，或者一行表示所有匹配。

您可以使用cross apply 生成行，如下所示：

SELECT 
    COALESCE(xl.ID, dw.ID) ID, 
    xl.name as excel_name,dw.name as dw_name,
    xl.flag as excel_flag,dw.flag as dw_flag,
    xl.city as excel_city,dw.city as dw_city,
    x.result
FROM source_excel xl
FULL OUTER JOIN source_dw dw ON xl.id = dw.id
CROSS APPLY (VALUES
    (CASE WHEN xl.ID IS NULL           THEN 'MISSING IN EXCEL' END),
    (CASE WHEN dw.ID IS NULL           THEN 'MISSING IN DW'    END),
    (CASE WHEN WHEN xl.NAME <> dw.NAME THEN 'NAME_MISMATCH'    END),
    (CASE WHEN xl.CITY <> dw.CITY      THEN 'CITY_MISMATCH'    END),
    (CASE WHEN xl.FLAG <> dw.FLAG      THEN 'FLAG_MISMATCH'    END),
    (CASE WHEN 
        xl.ID = dw.ID 
        AND xl.NAME = dw.NAME
        AND xl.CITY = dw.CITY
        AND xl.FLAG = dw.FLAG
    THEN 'ALL_MATCH' END)
) x(result)
WHERE x.result IS NOT NULL

【讨论】：