这里是一个纯粹的MySQL的答案:
我已经扩展您的样本数据:
DROP TABLE IF EXISTS test;
CREATE TABLE test (
id INT(11) AUTO_INCREMENT PRIMARY KEY,
col_1 VARCHAR(128),
col_2 VARCHAR(128),
col_3 VARCHAR(128),
col_4 VARCHAR(128)
);
INSERT INTO test VALUES (1,'a','b','c','d'),(2,'a','b','d','c'),(3,'a','d','c','b'),(4,'a','c','d','b'),(5,'e','f','g','h'),(6,'f','h','g','e')
;
SELECT * FROM test;
+----+-------+-------+-------+-------+
| id | col_1 | col_2 | col_3 | col_4 |
+----+-------+-------+-------+-------+
| 1 | a | b | c | d |
| 2 | a | b | d | c |
| 3 | a | d | c | b |
| 4 | a | c | d | b |
| 5 | e | f | g | h |
| 6 | f | h | g | e |
+----+-------+-------+-------+-------+
一个辅助表开始发挥作用:
DROP TABLE IF EXISTS tmp_test;
CREATE TABLE tmp_test (id int, col varchar(128));
INSERT INTO tmp_test
SELECT id, col_1 FROM test
UNION
SELECT id, col_2 FROM test
UNION
SELECT id, col_3 FROM test
UNION
SELECT id, col_4 FROM test
;
然后我截断原始表
TRUNCATE TABLE test;
ALTER TABLE test AUTO_INCREMENT = 1;
,以便可以用“不同”值重新填充:
INSERT INTO test (col_1, col_2, col_3, col_4)
SELECT SUBSTRING_INDEX(gc, '@', 1), SUBSTRING(SUBSTRING_INDEX(gc, '@', 2) FROM LOCATE('@', gc) + 1), SUBSTRING(SUBSTRING_INDEX(gc, '@', 3) FROM LENGTH(SUBSTRING_INDEX(gc, '@', 3))), REVERSE(SUBSTRING_INDEX(REVERSE(gc), '@', 1))
FROM (
SELECT DISTINCT
GROUP_CONCAT(col ORDER BY col SEPARATOR '@') AS gc
FROM tmp_test
GROUP BY id
) sq;
SELECT * FROM test;
+----+-------+-------+-------+-------+
| id | col_1 | col_2 | col_3 | col_4 |
+----+-------+-------+-------+-------+
| 1 | a | b | c | d |
| 2 | e | f | g | h |
+----+-------+-------+-------+-------+
因此,所有3行都将被删除? – Mihai
请定义重复的集合,列的问题?重复项必须位于同一列吗? – AgeDeO
你会保留哪一个?什么是可能的值范围? – Strawberry