Dataset metadata when using aggregator

[shaimil]

Can anyone explain the following.

I'm loading some numeric data into a dataset. Datatypes are integer.

I use an aggregator prior to the ds which throws out a dfloat, which I then modify.

Without the aggregator the ds is created with int32 datatypes. However, with the aggregator the datatype is dfloat for the calculated field.

I have specified integer types in the column defs of the ds.

Regards
Shai

[ray.wurlod]

It's just how it works - an engineering solution to the possibility of, and need to prevent - integer overflow from occurring.