Question

我的数据库中有一个表，该表具有一组位值列。我想在逗号分隔的字符串中插入新列，具体取决于这些位列的值。

让我们说一下这些列的名称：

c1,c2,c3,c4

如果一列中的值等于1，我想包含映射的字符串，如下所示：

c1: 'campaign1'
c2: 'campaign2'
c3: 'campaign3'
c4: 'campaign4'

因此，例如，如果一行包含以下值：

c1 = 1, c2 = 0, c3 = 1, c4 = 0

我想在名为“广告系列”的新列中插入以下内容

'campaign1,campaign3'

关于如何实现此目标的任何建议？

Answer 1

我将使用stuff()和一些字符串操作来做到这一点：

select t.*,
       stuff( ((case when c1 = 1 then ',campaign1' else '' end) +
               (case when c2 = 1 then ',campaign2' else '' end) +
               (case when c3 = 1 then ',campaign3' else '' end) +
               (case when c4 = 1 then ',campaign4' else '' end)
              ), 1, 1, ''
            ) as campaigns
from t;

Answer 2

这可行，但可能不是最佳解决方案：

DECLARE @table TABLE (id INT, c1 INT, c2 INT, c3 INT, c4 INT);
INSERT INTO @table 
SELECT 1, 1, 0, 1, 0
UNION ALL
SELECT 2, 0, 0, 1, 1;
WITH cte AS (
    SELECT
        id,
        CASE WHEN c1 = 1 THEN 'campaign1,' ELSE '' END + 
        CASE WHEN c2 = 1 THEN 'campaign2,' ELSE '' END + 
        CASE WHEN c3 = 1 THEN 'campaign3,' ELSE '' END + 
        CASE WHEN c4 = 1 THEN 'campaign4,' ELSE '' END AS campaigns
    FROM
        @table)
SELECT
    id,
    CASE 
        WHEN LEN(campaigns) > 0 THEN LEFT(campaigns, LEN(campaigns) - 1)
        ELSE ''
    END AS campaigns
FROM
    cte;

Gordon击败了我使用STUFF方法，当我意识到这种方法更好时，我匆忙编写了该方法，但这是一个答案，它也显示了UPDATE可能如何工作，因为这是您最初的方法要求：

DECLARE @table TABLE (id INT, c1 INT, c2 INT, c3 INT, c4 INT, campaigns VARCHAR(512));
INSERT INTO @table 
SELECT 1, 1, 0, 1, 0, NULL
UNION ALL
SELECT 2, 0, 0, 1, 1, NULL;
UPDATE
    t
SET
    campaigns = STUFF(CASE WHEN c1 = 1 THEN ',campaign1' ELSE '' END + 
        CASE WHEN c2 = 1 THEN ',campaign2' ELSE '' END + 
        CASE WHEN c3 = 1 THEN ',campaign3' ELSE '' END + 
        CASE WHEN c4 = 1 THEN ',campaign4' ELSE '' END, 1, 1, '')
FROM
    @table t;

SELECT * FROM @table;

结果：

id  c1  c2  c3  c4  campaigns
1   1   0   1   0   campaign1,campaign3
2   0   0   1   1   campaign3,campaign4

Answer 3

您可以使用IIF函数（前提是您使用的是 SQL Server 2012 + ），如下所示：

select IIF(c1=1, 'Campaign1', '')+IIF(c2=1 and c1>0,',','')+
       IIF(c2=1, 'Campaign2', '')+IIF(c3=1 and c1+c2>0,',','')+
       IIF(c3=1, 'Campaign3', '')+IIF(c4=1 and c1+c2+c3>0,',','')+
       IIF(c4=1, 'Campaign4', '') as 'Campaigns'
  from tab t;

SQL Fiddle Demo

Answer 4

只需给出另一个选择，您可以使用concat并替换掉

DECLARE @table TABLE (id INT, c1 INT, c2 INT, c3 INT, c4 INT)

INSERT INTO @table 
SELECT 1, 1, 0, 1, 0 UNION ALL SELECT 2, 0, 0, 1, 1

select t.*,
       replace(replace(c1, '1', 'Campaing1,'), '0', '') + 
       replace(replace(c2, '1', 'Campaing2,'), '0', '') + 
       replace(replace(c3, '1', 'Campaing3,'), '0', '') + 
       replace(replace(c4, '1', 'Campaing4,'), '0', '') 
from   @table t

结果是

id  c1  c2  c3  c4  COLUMN1 
--  --  --  --  --  ------- 
1   1   0   1   0   Campaing1,Campaing3,    
2   0   0   1   1   Campaing3,Campaing4,

但是我会去找戈登的答案

根据现有列选择逗号分隔的字符串

4 个答案: