/**
* Change the algebriac function type for algebraic functions in map and combine In map and
* combine the algebraic functions will be leaf of the plan
*
* @param fe
* @param type
* @throws PlanException
*/
private static void changeFunc(POForEach fe, byte type) throws PlanException {
for (PhysicalPlan plan : fe.getInputPlans()) {
List<PhysicalOperator> leaves = plan.getLeaves();
if (leaves == null || leaves.size() != 1) {
int errCode = 2019;
String msg = "Expected to find plan with single leaf. Found " + leaves.size() + " leaves.";
throw new PlanException(msg, errCode, PigException.BUG);
}
PhysicalOperator leaf = leaves.get(0);
if (leaf instanceof POProject) {
continue;
}
if (!(leaf instanceof POUserFunc)) {
int errCode = 2020;
String msg =
"Expected to find plan with UDF or project leaf. Found "
+ leaf.getClass().getSimpleName();
throw new PlanException(msg, errCode, PigException.BUG);
}
POUserFunc func = (POUserFunc) leaf;
try {
func.setAlgebraicFunction(type);
} catch (ExecException e) {
int errCode = 2075;
String msg = "Could not set algebraic function type.";
throw new PlanException(msg, errCode, PigException.BUG, e);
}
}
}
@Override
public void visitProject(POProject proj) throws VisitorException {
// check if this project is preceded by PODistinct and
// has the return type bag
List<PhysicalOperator> preds = mPlan.getPredecessors(proj);
if (preds == null) return; // this is a leaf project and so not interesting for patching
PhysicalOperator pred = preds.get(0);
if (preds.size() == 1 && pred instanceof PODistinct) {
if (distinct != null) {
// we should not already have been patched since the
// Project-Distinct pair should occur only once
int errCode = 2076;
String msg =
"Unexpected Project-Distinct pair while trying to set up plans for use with combiner.";
throw new OptimizerException(msg, errCode, PigException.BUG);
}
// we have stick in the POUserfunc(org.apache.pig.builtin.Distinct)[DataBag]
// in place of the Project-PODistinct pair
PhysicalOperator distinctPredecessor = mPlan.getPredecessors(pred).get(0);
POUserFunc func = null;
try {
String scope = proj.getOperatorKey().scope;
List<PhysicalOperator> funcInput = new ArrayList<PhysicalOperator>();
FuncSpec fSpec = new FuncSpec(DISTINCT_UDF_CLASSNAME);
funcInput.add(distinctPredecessor);
// explicitly set distinctPredecessor's result type to
// be tuple - this is relevant when distinctPredecessor is
// originally a POForeach with return type BAG - we need to
// set it to tuple so we get a stream of tuples.
distinctPredecessor.setResultType(DataType.TUPLE);
func =
new POUserFunc(
new OperatorKey(scope, NodeIdGenerator.getGenerator().getNextNodeId(scope)),
-1,
funcInput,
fSpec);
func.setResultType(DataType.BAG);
mPlan.replace(proj, func);
mPlan.remove(pred);
// connect the the newly added "func" to
// the predecessor to the earlier PODistinct
mPlan.connect(distinctPredecessor, func);
} catch (PlanException e) {
int errCode = 2077;
String msg = "Problem with reconfiguring plan to add distinct built-in function.";
throw new OptimizerException(msg, errCode, PigException.BUG, e);
}
distinct = func;
}
}