The current SchemaScanner doesn't seem to respect data types set in the upstream transformer - e.g. AttributeManager etc.
For example, an attribute contains 5 digits code and you would like to write the attribute values with “varchar(5)” into the destination dataset, since the values are integer mostly, but non-number characters could appear in some source datasets.
The current SchemaScanner determines its data type as a numeric type when there are number only in the attribute values, even if you have set "varchar(5)" to its data type in the upstream AttributeManager.
It would be great if SchemaScanner has an option to adopt data types which the user explicitly set in the upstream transformer.
