Wrangler allows to have incorrect schema and non-matching columns are populated with no data.
CDAP wrangler doesn't fail(Or produces validation error) if the output schema is different. For the non-matching columns, as expected, no data is populated.
This is detrimental at times when manual schema enforcement is not an option. e.g input source with a large number of fields (e.g. files with 100 or more columns)