In the workspace, attributes have their own datatype which is sometimes different from the original data type (in the source).
For exemple, a "date attribute" in the datasource is automaticaly converted into a "string attribute" into the workspace.
Really anoying for Python scripts to work on these attributes and to detect the type of the data without auditing every value of the attribute.
In this case, it is nearly impossible to create a dynamic schema with date columns without auditing every values.
It would be really easier to work with original datatype.
