当过滤器是逗号分隔的值列表时如何过滤记录

Question

我有一个表，其中有一列包含一串逗号分隔值 (CSV)。

例如VALUE1、VALUE2、VALUE3

作为过滤器传入的是另一组 CSV。

例如VALUE2，VALUE3

所以在上面的例子中，查询应该返回任何过滤值在 CSV 列中的记录。

示例

declare @table table
(
    rownum int,
    csv nvarchar(300)
)

insert into @table values (1,'VALUE1, VALUE2, VALUE3')
insert into @table values (2,'VALUE1, VALUE2')
insert into @table values (3,'VALUE1, VALUE3')
insert into @table values (4,'VALUE3, VALUE4')
insert into @table values (5,'VALUE1, VALUE2, VALUE3')
insert into @table values (6,'VALUE3, VALUE4, VALUE2')
insert into @table values (7,'VALUE3')

declare @Filter nvarchar(50)

set @Filter = 'VALUE1,VALUE2'

select * from @table

因此，在上面的示例中，第 1、2、3、5 和 6 行应该由查询返回，因为它们都包含 VALUE1 或 VALUE2。

Answer 1

如果我很好理解，这将解决它：

您创建一个函数来进行拆分：

CREATE FUNCTION [dbo].[ufn_CSVToTable] ( @StringInput VARCHAR(8000), @Delimiter nvarchar(1))
RETURNS @OutputTable TABLE ( [String] VARCHAR(10) )
AS
BEGIN

    DECLARE @String    VARCHAR(10)

    WHILE LEN(@StringInput) > 0
    BEGIN
        SET @String      = LEFT(@StringInput, 
                                ISNULL(NULLIF(CHARINDEX(@Delimiter, @StringInput) - 1, -1),
                                LEN(@StringInput)))
        SET @StringInput = SUBSTRING(@StringInput,
                                     ISNULL(NULLIF(CHARINDEX(@Delimiter, @StringInput), 0),
                                     LEN(@StringInput)) + 1, LEN(@StringInput))

        INSERT INTO @OutputTable ( [String] )
        VALUES ( @String )
    END

    RETURN
END
GO

那么你可以这样做：

declare @table table
(
    rownum int,
    csv nvarchar(300)
)

insert into @table values (1,'VALUE1, VALUE2, VALUE3')
insert into @table values (2,'VALUE1, VALUE2')
insert into @table values (3,'VALUE1, VALUE3')
insert into @table values (4,'VALUE2, VALUE3')
insert into @table values (5,'VALUE1, VALUE2, VALUE3')
insert into @table values (6,'VALUE3, VALUE1, VALUE2')
insert into @table values (7,'VALUE2, VALUE1')

declare @Filter nvarchar(50)

set @Filter = 'VALUE3,VALUE4'

select * from @table
SELECT * INTO #FilterTable FROM ufn_CSVToTable(@Filter, ',')
SELECT * FROM #FilterTable
select * from @table where EXISTS(SELECT String FROM #FilterTable WHERE csv like '%' + String + '%')
DROP TABLE #FilterTable

我正在考虑描述“返回包含过滤器中任何计算的任何行”

Answer 2

您可以使用 PatIndex 找出该字段中的模式。像下面这样：

SELECT * FROM @TABLE WHERE PATINDEX('%'+@FILTER+'%',CSV)>0

Answer 3

现在呢？我根本没有使用任何用户定义的函数。

declare @table table
(
    rownum int,
    csv nvarchar(300)
)

insert into @table values (1,'VALUE1,VALUE2,VALUE3')
insert into @table values (2,'VALUE1,VALUE2')
insert into @table values (3,'VALUE1,VALUE3')
insert into @table values (4,'VALUE2,VALUE3')
insert into @table values (5,'VALUE1,VALUE2,VALUE3')
insert into @table values (6,'VALUE3,VALUE1,VALUE2')
insert into @table values (7,'VALUE2,VALUE1')

DECLARE @FILTER VARCHAR(50)
DECLARE @IN VARCHAR(MAX)
SET @FILTER='VALUE1,VALUE2'
SET @IN=REPLACE(@FILTER,',','%')
SET @IN='%'+@IN+'%'
SELECT @IN
SELECT * FROM @table WHERE PATINDEX(@IN,RTRIM(LTRIM(CSV))) >0

Answer 4

试试这个 -

DECLARE @table TABLE (rownum INT, csv NVARCHAR(300))

INSERT INTO @table 
VALUES 
       (1,'VALUE1, VALUE2, VALUE3')
     , (2,'VALUE1, VALUE2')
     , (3,'VALUE1, VALUE3')
     , (4,'VALUE2, VALUE3')
     , (5,'VALUE1, VALUE2, VALUE3')
     , (6,'VALUE3, VALUE1, VALUE2')
     , (7,'VALUE2, VALUE1')

DECLARE @Filter NVARCHAR(50)
SELECT @Filter = 'VALUE1,VALUE2'

;WITH cte AS 
(
     SELECT token = SUBSTRING(
          t.string
        , n.number + 1
        , ABS(CHARINDEX(',', t.string, n.number + 1) - n.number - 1))
     FROM ( SELECT string = ',' + @Filter ) t
     CROSS JOIN [master].dbo.spt_values n
     WHERE n.[type] = 'p'
         AND n.number <= LEN(t.string)
         AND SUBSTRING(t.string, n.number, 1) = ','
)
SELECT DISTINCT rownum, csv
FROM cte
JOIN @table ON PATINDEX('%'+ token +'%', CSV) = 0