[PIG-3265] Pig uses Python UDF registered in the last name space

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.9.2, 0.10.0, 0.11
• Component/s: impl
• Assignee: Unassigned
• Reporter: Viraj Bhat
• Created: 01/Apr/13 21:24
• Updated: 01/Apr/13 21:27

Description

Pig uses Python UDF registered in the last name space as you can see below using a Pig script like this:

-- test_jython.pig --
register 'B.py' using org.apache.pig.scripting.jython.JythonScriptEngine as B;
register 'A.py' using org.apache.pig.scripting.jython.JythonScriptEngine as A;

I = load 'test.txt' as v;
O = foreach I generate A.a(v) as a, A.b(v) as b;
dump O;
-- END --

A.py

@outputSchema("word:int")
def a(v):
    return 'a

B.py

@outputSchema("word1:int")
def b(v):
    return 'b'

data.txt

1
2
3

2013-04-01 20:17:36,338 [main] INFO org.apache.pig.scripting.jython.JythonScriptEngine - Register scripting UDF: B.b
2013-04-01 20:17:36,377 [main] INFO org.apache.pig.scripting.jython.JythonScriptEngine - Register scripting UDF: A.b
2013-04-01 20:17:36,378 [main] INFO org.apache.pig.scripting.jython.JythonScriptEngine - Register scripting UDF: A.a

The class JythonScriptEngine.java, and function: registerFunctions(), a hashmap in PigContext is populated with latest namespace and function. This needs to be changed and a check should be conducted before populating the hashmap. A unit test also needs to be written.

This seems to be related to https://issues.apache.org/jira/browse/PIG-2404 which was closed as INVALID
Viraj

Issue Links

• is related to: PIG-2404 NullPointerException when I have multiple python udfs