Try this 🙂 It works for me on string type date columns, holding something like this inside: 2016-05-02T18:28:15.790+0000

df = df1.select("some_id", year(df1["date"]).alias('year'), month(df1["date"]).alias('month'), dayofmonth(df1["date"]).alias('day'), hour(df1["date"]).alias('hour')).show()