Question

我有一个导入过程，需要更新表中旧记录的“结束日期”。导入来自其他人，我需要一些与特定用户信息匹配的细节。我需要做的是更新旧记录，在下面的示例中，该记录将是UserID = 1，而不是更新UserID = 4。

我有以下sql要更新，但正如你所知，它会更新两个记录：

UPDATE t1
SET t1.[EndDate] = t2.[EndDate]
FROM ExistingUser AS t1, ImportUser AS t2
WHERE (t2.[uName] = t1.[uName]) AND (t1.[EndDate] IS NULL);

免责声明：我没有创建数据库，也无法重新设计表格，所以请怜悯我。感谢!!!

ExistingUser - 表

UserID    uName   BeginDate    EndDate
  1       John   01/01/2013
  2       Mary   05/01/2014   04/30/2015
  3       Bob    12/01/2014
  4       John   06/01/2015

ImportUser - 表

uName    EndDate
John    05/31/2015

Answer 1

这可以通过构建一个子查询来完成，该子查询使用没有填充EndDate的用户名对条目进行编号。这个子查询看起来像这样：

SELECT [UserID], [uName],
  ROW_NUMBER() OVER (PARTITION BY [uName] ORDER BY [BeginDate]) AS IX
FROM ExistingUser
WHERE [EndDate] IS NULL

演示：http://www.sqlfiddle.com/#!6/72c49/2

| UserID | uName | IX |
|--------|-------|----|
|      3 |   Bob |  1 |
|      1 |  John |  1 |
|      4 |  John |  2 |

在此子查询中，具有相同uName John的两个记录按BeginDate顺序编号。现在，如果我们将原始UPDATE查询加入此子查询中搜索IX=1的记录，我们将只更新最早的行。

WITH EarliestExistingUser AS (
  SELECT [UserID], [uName],
    ROW_NUMBER() OVER (PARTITION BY [uName] ORDER BY [BeginDate]) AS IX
  FROM ExistingUser
  WHERE [EndDate] IS NULL
)
UPDATE t1
SET t1.[EndDate] = t2.[EndDate]
FROM ExistingUser AS t1
  JOIN EarliestExistingUser AS t1f ON t1.UserID = t1f.UserID
,ImportUser AS t2
WHERE (t2.[uName] = t1.[uName]) AND (t1.[EndDate] IS NULL)
AND (t1f.IX = 1)

演示：http://www.sqlfiddle.com/#!6/72c49/3

SQL Update第一个记录实例--SQL Server

1 个答案: