Re: Custom UDF + Grouping - Unexpected Output

[Daniel Dai]

In your udf:
if (num1 < num2)
{
   t.set(0, num1 + "," + num2);
}
else if (num2 < num1)
{
   t.set(0, num2 + "," + num1);
}
You actually only put one item into the tuple. So your UDF generate a 
bag of tuples of one field, not two fields.
I think what you mean is:
if (num1 < num2)
{
   t.set(0, num1);
   t.set(1, num2);
}
else if (num2 < num1)
{
   t.set(0, num2);
   t.set(1, num1);
}

Daniel

Michael Moss wrote:
Thanks, Daniel.

My UDF with schema (which I suspect is culprit) is below. I've tried
excluding the "outputSchema()" method entirely and a several variations:

(Full source here: http://pastie.org/1362084)

public class NormalizeListUDF extends EvalFunc<DataBag>
{
public DataBag exec(Tuple input) throws IOException
{
if (input == null || input.size() == 0)
return null;
try
{
DataBag output = DefaultBagFactory.getInstance().newDefaultBag();

List<Object> tuples = input.getAll();
String line = (String) tuples.remove(0);
line = line.trim();
String[] items = line.split(",");

for (int i = 1; i < items.length - 1; i++)
{
for (int j = i + 1; j < items.length; j++)
{
int num1 = Integer.parseInt(items[i]);
int num2 = Integer.parseInt(items[j]);

Tuple t = TupleFactory.getInstance().newTuple(1);

if (num1 < num2)
{
t.set(0, num1 + "," + num2);
}
else if (num2 < num1)
{
t.set(0, num2 + "," + num1);
}
output.add(t);
}
}
return output;
}
catch (Exception e)
{
throw WrappedIOException.wrap("Caught exception processing input row ", e);
}
}

public Schema outputSchema(Schema input)
{
try
{
List<Schema.FieldSchema> fields = new ArrayList<Schema.FieldSchema>();
Schema.FieldSchema f1 = new Schema.FieldSchema("f1", DataType.INTEGER);
Schema.FieldSchema f2 = new Schema.FieldSchema("f2", DataType.INTEGER);
fields.add(f1);
fields.add(f2);
 Schema tupleInner = new Schema(fields);
Schema.FieldSchema tupleSchema = new Schema.FieldSchema("t1", tupleInner,
DataType.TUPLE);

Schema bagInner = new Schema(tupleSchema);
Schema.FieldSchema bagSchema = new Schema.FieldSchema("bag", bagInner,
DataType.BAG);
return new Schema(bagSchema);
}
catch (Exception e)
{
return null;
}
}
}

On Wed, Dec 8, 2010 at 7:04 PM, Daniel Dai <jiany...@yahoo-inc.com> wrote:

  
It is not expected. I would think something wrong inside NormalizeListUDF.
Make sure you feed bag of tuples which has the schema (int, int) inside your
UDF. If you can post your UDF, I can know better.

Daniel


Michael Moss wrote:

    
Hello,

I'm having an issue with a script that uses an EvalFunc I wrote. The issue
is the final output contains characters that I am not expecting (commas -
followed by what I'm guessing are null fields which I do not see).

Snippet:
C = FOREACH B GENERATE FLATTEN(B) as (f1:int,f2:int);
grunt> DUMP C;
(2,3)
(2,4)
(2,5)
(3,4)
(3,5)
(4,5)
(2,3)
(2,4)
(2,5)
(3,4)
(3,5)
(4,5)

D = GROUP C by (f1,f2);
grunt> describe D;
D: {group: (f1: int,f2: int),C: {f1: int,f2: int}}

grunt> DUMP D;
((2,3,),{(2,3,),(2,3,)})
((2,4,),{(2,4,),(2,4,)})
((2,5,),{(2,5,),(2,5,)})
((3,4,),{(3,4,),(3,4,)})
((3,5,),{(3,5,),(3,5,)})
((4,5,),{(4,5,),(4,5,)})

My question is, what are these extra comma/null fiends in each tuple? I
expected the first row to read as:
((2,3),{(2,3),(2,3)})

It seems related, but when I run 'ILLUSTRATE C', I get an exeption:
java.lang.IndexOutOfBoundsException: Index: 1, Size: 1
at java.util.ArrayList.RangeCheck(ArrayList.java:547)
at java.util.ArrayList.get(ArrayList.java:322)
at org.apache.pig.data.DefaultTuple.get(DefaultTuple.java:143)
at org.apache.pig.pen.util.ExampleTuple.get(ExampleTuple.java:80)
at

org.apache.pig.pen.util.DisplayExamples.MakeArray(DisplayExamples.java:190)
at

org.apache.pig.pen.util.DisplayExamples.printTabular(DisplayExamples.java:86)
at

org.apache.pig.pen.util.DisplayExamples.printTabular(DisplayExamples.java:69)
at
org.apache.pig.pen.ExampleGenerator.getExamples(ExampleGenerator.java:143)
at org.apache.pig.PigServer.getExamples(PigServer.java:785)
at

org.apache.pig.tools.grunt.GruntParser.processIllustrate(GruntParser.java:555)
at

org.apache.pig.tools.pigscript.parser.PigScriptParser.parse(PigScriptParser.java:246)
at

org.apache.pig.tools.grunt.GruntParser.parseStopOnError(GruntParser.java:162)
at

org.apache.pig.tools.grunt.GruntParser.parseStopOnError(GruntParser.java:138)
at org.apache.pig.tools.grunt.Grunt.run(Grunt.java:75)
at org.apache.pig.Main.main(Main.java:357)

Excruciating detail below:

My script:
REGISTER udf.jar
A = LOAD '/pig_input/co.txt' as (line:chararray);
B = FOREACH A GENERATE com.thumbplay.pig.NormalizeListUDF(line) as B;
C = FOREACH B GENERATE FLATTEN(B) as (f1:int,f2:int);
D = GROUP C by (f1,f2);
E = FOREACH D GENERATE group, COUNT(C);
STORE E INTO 'output' USING PigStorage(',');

Here's what I'm trying to do:
For input:
A,1,2,3
B,1,2,3

Produce combinations for each row (My UDF does this):
(1,2),(1,3),(2,3)
(1,2),(1,3),(2,3)

Flatten them:
(1,2),
(1,3),
(2,3),
(1,2),
(1,3),
(2,3)

Group and count them:
(1,2),2
(1,3),2
(2,3),2