起因是公司的crm录入不规范，有重复数据。

之后考虑到需要手动处理，首先需要自动找出重复的数据

查重要求：

存在多个不允许重复的字段，任一字段重复，则判断为同一个客户。划分到同一重复组中。

查重sql如下

SELECT CONCAT('组', dense_rank() OVER (ORDER BY group_key)) AS group_info,id, name, tel
FROM (SELECT id, name, tel,CONCAT_WS('-', IF(name IS NULL, '', name),  IF(tel IS NULL, '', '1')) AS group_key,COUNT(*) OVER (PARTITION BY CONCAT_WS('-', IF(name IS NULL, '', name),  IF(tel IS NULL, '', '1'))) AS group_countFROM customer
) AS subquery
WHERE group_count > 1
ORDER BY group_key

通过上述sql可以获得一个查重的sql。

那么接下来的问题是如果重复数据大，我们是需要一个分页的。首先我们需要知道重复了多少组。

用以下sql获取：

SELECT COUNT(*) AS total_count 
FROM (SELECT DISTINCT group_infoFROM (SELECT CONCAT('组', dense_rank() OVER (ORDER BY group_key)) AS group_info,id, name, tel
FROM (SELECT id, name, tel,CONCAT_WS('-', IF(name IS NULL, '', name),  IF(tel IS NULL, '', '1')) AS group_key,COUNT(*) OVER (PARTITION BY CONCAT_WS('-', IF(name IS NULL, '', name),  IF(tel IS NULL, '', '1'))) AS group_countFROM customer
) AS subquery
WHERE group_count > 1
ORDER BY group_key) AS data_counts
) AS distinct_groups;

这里的分页应当按照重复组分页，传统的分页就失效了。所以我们就需要手动计算分页。

把分组字段group_info中的组字去掉，按group_info进行大于，小于判断

SELECT dense_rank() OVER (ORDER BY group_key)  AS group_info,id, name, tel
FROM (SELECT id, name, tel,CONCAT_WS('-', IF(name IS NULL, '', name),  IF(tel IS NULL, '', '1')) AS group_key,COUNT(*) OVER (PARTITION BY CONCAT_WS('-', IF(name IS NULL, '', name),  IF(tel IS NULL, '', '1'))) AS group_countFROM customer
) AS subquery
WHERE group_count > 1
AND group_info>0 AND group_info<11
ORDER BY group_key