spark dataset根据某行数据一行变多行

用spark sql语句就能实现。

select * from t1  lateral view explode(split(label_class_tmp, ',')) tmpTable as label_class