jython udf support null

[allwefantasy]

What changes were proposed in this pull request?

• Finshed changes describe

How was this patch tested?

• Testing done

Are there and DOC need to update?

• Doc is finished

Spark Core Compatibility

set convert_data='''

def apply(self,m):
    return "data:" + str(m)
''';

load script.`convert_data` as scriptTable;

register ScriptUDF.`scriptTable` as convert_data
options lang="python" and dataType="string";

select convert_data(null) as a ,"c" as c as b;