@Pien Derkx​ Hi, It would be really helpful if you can share some sample values that you are trying to typecast along with your expected result so that we can try that. But I have seen that in the new Spark versions, you can just use a single literal for the date formats Ex MM/dd/yyyy --> m/d/y. This will work. Please see if you can try the same like uwy and let me know.