背景：存数据的时候多存了，导致数据重复

思路：group by 对重复的字段进行分组，重复意味着group by count的结果大于1，将count大于1的字段的最小id保留，其余删除

语句

delete from base_cell_extend where   data_category =6 and data_key='device_num' and 
cgi in (select cgi from base_cell_extend where  data_category =6 and data_key='device_num' group by cgi having count(cgi) > 1 )
and base_cell_extend.id not in (select min(base_cell_extend.id) from base_cell_extend where  data_category =6 and data_key='device_num' group by cgi having count(cgi) > 1 );

注意：最好还是能够在删除之前先select，并且计算一下删除后的数据总数是否正确，再进行delete