[PIG-2031] NPE in TOP

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Labels: newbie
• Assignee: Jacob Perkins
• Reporter: Jacob Perkins
• Created: 03/May/11 13:29
• Updated: 12/Jul/15 13:57
• Resolved: 12/Jul/15 13:57

Description

If a NULL DataBag is passed to org.apache.pig.builtin.TOP then a NPE is thrown. Consider:

$: cat foo.tsv
a  {(foo,1),(bar,2)}
b
c  {(fyha,4),(asdf,9)}

then:

data  = LOAD 'foo.tsv' AS (key:chararray, a_bag:bag {t:tuple (name:chararray, value:int)});
tpd   = FOREACH data {
          top_n = TOP(1, 1, a_bag);
          GENERATE
            key   AS key,
            top_n AS top_n
          ;             
        };
DUMP tpd;

will throw an NPE when it gets to the row with no bag.