Question

我忘了在表中添加唯一索引（process_id，result_id）：

id, created_at, process_id, result_id, value
1, 2018-11-01, 1, 1, a
2, 2018-11-03, 1, 1, b
3, 2018-11-05, 2, 4, c

现在，对于process_id和result_id的相同组合，我有多个条目。

为了添加唯一索引，我需要消除重复项，并且只保留每个过程/结果对的最新条目（created_at）。

如何轻松识别和删除较旧的重复项？

我可以识别所有重复项：

select pr.id, pr.created_at, prd.* 
from process_results pr
join (
    select process_id, result_id
    from process_results 
    group by process_id, result_id
    having count(*) > 1
) prd on pr.request_id = prd.request_id and pr.request_filter_id = prd.request_filter_id

但我不确定如何删除除最新的所有内容

预期结果：

id, created_at, process_id, result_id, value
2, 2018-11-03, 1, 1, b
3, 2018-11-05, 2, 4, c

Answer 1

DELETE t1 
FROM yourTable t1
INNER JOIN yourTable t2 
ON
    t1.created_at < t2.created_at 
AND t1.process_id = t2.process_id 
AND t1.result_id = t2.result_id    
;

识别相同

SELECT t1.*
FROM yourTable t1
INNER JOIN yourTable t2 
ON
    t1.created_at < t2.created_at 
AND t1.process_id = t2.process_id 
AND t1.result_id = t2.result_id

Answer 2

DELETE bad_rows.* FROM process_results AS bad_rows
                   INNER JOIN (
                      SELECT created_at,value,min(`process_id`) AS min_id FROM process_results GROUP BY process_id,result_id HAVING COUNT(*) > 1
                   ) AS good_rows ON good_rows.value = bad_rows.value AND good_rows.min_id != bad_rows.id

删除“重复的”条目并保留最新的

2 个答案: