[PIG-4756] Groovy algebraic UDF only works when return type is long

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Assignee: Daniel Dai
• Reporter: Daniel Dai
• Created: 06/Dec/15 10:26
• Updated: 07/Dec/15 02:34

Description

The following groovy UDF does not work:

import org.apache.pig.scripting.groovy.AlgebraicInitial;
import org.apache.pig.scripting.groovy.AlgebraicIntermed;
import org.apache.pig.scripting.groovy.AlgebraicFinal;
import java.lang.Math;

class GroovyUDFs {
  @AlgebraicInitial('max')
  public static Tuple initial(Tuple t) {
    return t[1].next();
  }
  @AlgebraicIntermed('max')
  public static Tuple intermed(Tuple t) {
    double max = -1;
    for (Tuple inner: t[1]) {
      max = Math.max(max, inner[0]);
    }
    return [max];
  }
  @AlgebraicFinal('max')
  public static double algFinal(Tuple t) {
    double max = -1;
    for (Tuple inner: t[1]) {
      max = Math.max(max, inner[0]);
    }
    return max;
  }
}

Error message:

Caused by: java.lang.RuntimeException: could not instantiate 'org.apache.pig.scripting.groovy.GroovyAlgebraicEvalFunc$DoubleGroovyAlgebraicEvalFunc$Final' with arguments 'null'
	at org.apache.pig.impl.PigContext.instantiateFuncFromSpec(PigContext.java:766)
	at org.apache.pig.EvalFunc.getReturnTypeFromSpec(EvalFunc.java:153)
	at org.apache.pig.EvalFunc.<init>(EvalFunc.java:144)
	at org.apache.pig.AccumulatorEvalFunc.<init>(AccumulatorEvalFunc.java:33)
	at org.apache.pig.AlgebraicEvalFunc.<init>(AlgebraicEvalFunc.java:82)
	at org.apache.pig.scripting.groovy.GroovyAlgebraicEvalFunc.<init>(GroovyAlgebraicEvalFunc.java:39)
	at org.apache.pig.scripting.groovy.GroovyAlgebraicEvalFunc$DoubleGroovyAlgebraicEvalFunc.<init>(GroovyAlgebraicEvalFunc.java:199)
	... 42 more
Caused by: java.lang.InstantiationException: org.apache.pig.scripting.groovy.GroovyAlgebraicEvalFunc$DoubleGroovyAlgebraicEvalFunc$Final
	at java.lang.Class.newInstance(Class.java:364)
	at org.apache.pig.impl.PigContext.instantiateFuncFromSpec(PigContext.java:736)
	... 48 more

It works when the return type is long.