Question

对标题感到抱歉，但这个问题难以理解。我正在处理一个构造不良的数据库，并遇到了一个在同一列上有多个记录的表，如下所示：

| Column Header |

Record1; Record2; Record3

我想询问SQL中是否有任何方法可以使用;作为分隔符来分隔这些记录。我问这个是因为重建数据库目前不是一种选择。

Answer 1

快速解决方案 -

DECLARE @T TABLE
(
    Col NVARCHAR(2000),
    Col1 AS PARSENAME(REPLACE(Col, '; ', '.'), 3),
    Col2 AS PARSENAME(REPLACE(Col, '; ', '.'), 2),
    Col3 AS PARSENAME(REPLACE(Col, '; ', '.'), 1)
)

INSERT INTO @T (Col)
VALUES ('Record1; Record2; Record3')

SELECT * FROM @T

良好的解决方案 - 规范化表

Answer 2

DECLARE @badData TABLE (id INT NOT NULL, txt NVARCHAR(max));

INSERT INTO @badData
VALUES (1, 'foo,bar,baz'), (2, NULL);;

-- the idea is to recursively 'pop' a value from the start of the string, splitting it into 'head' and 'tail' components
WITH unpacked (id, head, tail)
AS (
  SELECT id, LEFT(txt, CHARINDEX(',', txt + ',') - 1), STUFF(txt, 1, CHARINDEX(',', txt + ','), '')
  FROM @badData

  UNION ALL

  SELECT id, LEFT(tail, CHARINDEX(',', TAIL + ',') - 1), STUFF(tail, 1, CHARINDEX(',', tail+ ','), '')
  FROM unpacked
  WHERE tail > ''
  )
SELECT id, head
FROM unpacked
ORDER BY id

严重抄袭https://stackoverflow.com/a/5493616/6722

分隔同一数据库列中的记录

2 个答案: