udf_annotation_error.py

@udf(returnType=ArrayType(StringType()))
def to_upper_list(s):
    return [i.upper() for i in s]

# Case 1 - UDF annotation
to_upper_list(['potato', 'carrot', 'tomato'])

"""
TypeError: Invalid argument, not a string or column: ['potato', 'carrot', 'tomato'] of type <class 'list'>. 
For column literals, use 'lit', 'array', 'struct' or 'create_map' function
"""

# Case 2 - UDF annotation and calling the method using the literal lit() method
to_upper_list(lit(['potato', 'carrot', 'tomato']))

"""
col = ['potato', 'carrot', 'tomato']

    def _(col):
        sc = SparkContext._active_spark_context
>       jc = getattr(sc._jvm.functions, name)(col._jc if isinstance(col, Column) else col)
E       AttributeError: 'NoneType' object has no attribute '_jvm'
"""