Question

我想要做的是相互匹配行，并排除所有匹配多次的行。

数据集（请记住，这是简化的，我需要从查询中检索更多信息）

表1：

NAME    PART1       PART2

ABC     FirstABC    SecondABC
BAC     FirstBAC    SecondBAC
BAC     SecondBAC   FirstBAC
CBA     FirstCBA    SecondCBA
DAE     FirstDAE    SecondDAE
EDA     FirstEDA    SecondEDA

表2：

ITEM1       ITEM2

FirstABC    SecondABC
FirstBAC    SecondBAC
FirstBAC    SecondBAC
FirstCBA    SecondCBA
FirstDAE    SecondDAE
FirstEDA    SecondEDA

到目前为止我所拥有的：

SELECT *
FROM TOMATCH as T
LEFT JOIN AGAINST as A
ON (
  T.Part1 = A.Item1
  OR T.Part2 = A.Item1
  OR T.Part1 = A.Item2
  OR T.Part2 = A.Item2
  )

我正在寻找什么结果：

NAME    PART1       PART2       ITEM1       ITEM2

ABC     FirstABC    SecondABC   FirstABC    SecondABC
CBA     FirstCBA    SecondCBA   FirstCBA    SecondCBA
DAE     FirstDAE    SecondDAE   FirstDAE    SecondDAE
EDA     FirstEDA    SecondEDA   FirstEDA    SecondEDA

Answer 1

您可以按所有字段进行分组，并添加一个count = 1的having子句。

SELECT name, part1, part2, item1, item2
FROM TOMATCH as T
LEFT JOIN AGAINST as A
ON (
 T.Part1 = A.Item1
 OR T.Part2 = A.Item1
 OR T.Part1 = A.Item2
 OR T.Part2 = A.Item2
 )
group by name, part1, part2, item1, item2
having count(*) = 1

请参阅SqlFiddle

Answer 2

在加入TOMATCH之前，将重复项从AGAINST中过滤掉。

UPDATE t
  SET UniqueMatch = ItemNumber
FROM TOMATCH t
INNER JOIN (
  SELECT Item1, Item2, MIN(ItemNumber) AS ItemNumber
  FROM AGAINST
  GROUP BY Item1, Item2
  HAVING COUNT(*) = 1
) a ON (Item1 = Part1 AND Item2 = Part2)

SQL显示不重复LEFT JOIN行

2 个答案: