datax同步mysql到doris，数据量多了一千多万

原表只有2600w多数据，抽取结束显示这么多。另外两个表倒是正常。
json配置没什么特别的，就是用了原表的id作为splitPk，配置了table,reader的column为*,数据库写在了jdbcUrl中。channel配置了4个，限速为-1。

Start by reproducing the MySQL-to-Doris sync with the reported JSON settings: id as splitPk, table and reader column set to *, four channels, and no rate limit. Compare source and destination row counts and inspect the configuration and logs to identify why roughly ten million extra rows are written; done means the counts match.

Written by the indexing model from the issue text.