add DynamicFrame.resolveChoice， redshift create new column type_short

I am using AWS Glue for data ETL to load data from MySQL into Redshift. In MySQL, the types of the three columns, type, sub_type, and status, are all unsigned int.

However, when stored in Redshift, they become NULL. After adding the method DynamicFrame.resolveChoice(specs=[('sub_type', 'cast:int'), ('type', 'cast:int'), ('status', 'cast:int')]), the problem of sub_type and status columns being NULL in Redshift is solved, but the problem of the type column being NULL persists. Now, there is an additional column, type_short, with a value of NULL. who can explain why this problem occurred and how to solve it?