Question

有一张表，该表包含97972561行（记录）和8列（属性）。格式如下：

+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+
| PREDICATE_ID | PMID | SENTENCE_ID | SUBJECT_ID | SUBJECT_NAME | PREDICATE | OBJECT_ID | OBJECT_NAME |
+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+

我想过滤仅主题，谓语和客体值仅出现一次的记录。例如，一个表中有四个记录。由于（Bob，is_a，Person）仅出现一次，因此应该从结果中排除最后一个记录。

+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+
| PREDICATE_ID | PMID | SENTENCE_ID | SUBJECT_ID | SUBJECT_NAME | PREDICATE | OBJECT_ID | OBJECT_NAME |
+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+
| 1            | 100  | 1           | 2          | Bob          | is_born_in| 3         | 1994        |
+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+
| 1            | 103  | 3           | 2          | Bob          | is_born_in| 3         | 1994        |
+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+
| 1            | 102  | 5           | 2          | Bob          | is_born_in| 3         | 1994        |
+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+
| 2            | 104  | 2           | 2          | Bob          | is_a      | 4         | Person      |
+--------------+------+-------------+------------+--------------+-----------+-----------+-------------+

任何帮助将不胜感激！

Answer 1

使用聚合，我们可以尝试：

SELECT t1.*
FROM yourTable t1
INNER JOIN
(
    SELECT SUBJECT_ID, PREDICATE_ID, OBJECT_ID
    FROM yourTable
    GROUP BY SUBJECT_ID, PREDICATE_ID, OBJECT_ID
    HAVING COUNT(*) > 1
) t2
    ON t1.SUBJECT_ID = t2.SUBJECT_ID AND
       t1.PREDICATE_ID = t2.PREDICATE_ID AND
       t1.OBJECT_ID = t2.OBJECT_ID;

如果您使用的是MySQL 8+，我们可以利用分析函数来生成更简洁的查询：

WITH cte AS (
    SELECT *, COUNT(*) OVER (PARTITION BY BY SUBJECT_ID, PREDICATE_ID, OBJECT_ID) cnt
    FROM yourTable
)

SELECT *
FROM cte
WHERE cnt > 1;

如何在mysql的多列中查找具有相同值的行

1 个答案: