Question

我想阅读公司表并从名称中取出所有可能的后缀。这是我到目前为止所做的：

declare @badStrings table (item varchar(50))

INSERT INTO @badStrings(item)
SELECT 'company' UNION ALL
SELECT 'co.' UNION ALL
SELECT 'incorporated' UNION ALL
SELECT 'inc.' UNION ALL
SELECT 'llc' UNION ALL
SELECT 'llp' UNION ALL
SELECT 'ltd'

select id, (companyname = Replace(name, item, '') FROM @badStrings)
from companies
where name != ''

Answer 1

Ed Northridge的回答是有效的，我已经对它进行了投票，但为了防止需要多次替换，我将使用他的示例数据添加另一个选项。例如，如果其中一家公司被称为“The PC Company LTD”，这将在输出中复制行，其中一行是“The PC LTD”，另一行是“The PC Company”。要解决此问题，根据您的预期结果，有2个选项。第一种是仅在名称末尾出现“Bad Strings”时替换它们。

SELECT  c.ID, RTRIM(x.Name) [Name]
FROM    @companies c
        OUTER APPLY 
        (   SELECT  REPLACE(c.name, item, '') AS [Name]
            FROM    @badStrings
                    -- WHERE CLAUSE ADDED HERE
            WHERE   CHARINDEX(item, c.Name) = 1 + LEN(c.Name) - LEN(Item)
        ) x
WHERE   c.name != '' 
AND     x.[Name] != c.Name

这将产生“PC公司”而没有重复。

另一个选项是递归地替换Bad Strings的所有出现：

;WITH CTE AS
(   SELECT  c.ID, c.Name [OriginalName], RTRIM(x.Name) [Name], 1 [Level]
    FROM    @companies c
            OUTER APPLY 
            (   SELECT  REPLACE(c.name, item, '') AS [Name]
                FROM    @badStrings
                WHERE   CHARINDEX(item, c.Name) = 1 + LEN(c.Name) - LEN(Item)
            ) x
    WHERE   c.name != '' 
    AND     RTRIM(x.Name) != c.Name
    UNION ALL
    SELECT  c.ID, OriginalName, RTRIM(x.Name) [Name], Level + 1 [Level]
    FROM    CTE c
            OUTER APPLY 
            (   SELECT  REPLACE(c.name, item, '') AS [Name]
                FROM    @badStrings
                WHERE   CHARINDEX(item, c.Name) = 1 + LEN(c.Name) - LEN(Item)
            ) x
    WHERE   c.name != '' 
    AND     x.[Name] != c.Name  
)

SELECT  DISTINCT ID, Name, OriginalName
FROM    (   SELECT  *, MAX(Level) OVER(PARTITION BY ID) [MaxLevel]
            FROM    CTE
        ) c
WHERE   Level = maxLevel

这将从“The PC Company”中获得“PC”。

Answer 2

我运行代码段的错误是：

Msg 102, Level 15, State 1, Line 12
Incorrect syntax near '='.

以下代码不是理想的解决方案 - 它只返回REPLACE函数更改了其名称已的公司列表。

declare @companies table (id int, name nvarchar(50))
INSERT INTO @companies(id, name)
SELECT 1,'One Company' UNION ALL
SELECT 2, 'Two co.' UNION ALL
SELECT 3, 'Three incorporated' UNION ALL
SELECT 4, 'Four inc.' UNION ALL
SELECT 5, 'Five llc' UNION ALL
SELECT 6, 'Six llp' UNION ALL
SELECT 7, 'Seven ltd'

select * from @companies

declare @badStrings table (item varchar(50))

INSERT INTO @badStrings(item)
SELECT 'company' UNION ALL
SELECT 'co.' UNION ALL
SELECT 'incorporated' UNION ALL
SELECT 'inc.' UNION ALL
SELECT 'llc' UNION ALL
SELECT 'llp' UNION ALL
SELECT 'ltd'

select * from @badStrings

以下是已编辑的查询：

select id, x.Name
from @companies c
OUTER APPLY (
    SELECT  Replace(c.name, item, '') AS [Name]
    FROM    @badStrings
) x    
where c.name != '' 
AND x.[Name] != c.Name

返回：

id          Name
----------- --------
1           One 
2           Two 
3           Three 
4           Four 
5           Five 
6           Six 
7           Seven 

(7 row(s) affected)

希望它有用

修改将匹配应用于结束且@badStrings值

的公司名称的替代方法
select id, x.Name from @companies c OUTER APPLY ( SELECT Replace(c.name, item, '') AS [Name] FROM @badStrings WHERE c.Name LIKE '%'+item ) x where c.name != ''

SQL多次替换

2 个答案: