Start line:  
End line:  

Snippet Preview

Snippet HTML Code

Stack Overflow Questions
Licensed to the Apache Software Foundation (ASF) under one or more contributor license agreements. See the NOTICE file distributed with this work for additional information regarding copyright ownership. The ASF licenses this file to you under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
 
 package org.apache.hadoop.hive.ql.udf.generic;
 
 
 import  org.apache.hadoop.hive.ql.exec.Description;
 import  org.apache.hadoop.hive.ql.exec.UDFArgumentTypeException;
 import  org.apache.hadoop.hive.ql.metadata.HiveException;
 import  org.apache.hadoop.hive.ql.parse.SemanticException;
 import  org.apache.hadoop.hive.serde2.io.DoubleWritable;
 import  org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector;
 import  org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory;
 import  org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector;
 import  org.apache.hadoop.hive.serde2.objectinspector.StructField;
 import  org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector;
 import  org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector;
 import  org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector;
 import  org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory;
 import  org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorUtils;
 import  org.apache.hadoop.hive.serde2.typeinfo.PrimitiveTypeInfo;
 import  org.apache.hadoop.hive.serde2.typeinfo.TypeInfo;
 import  org.apache.hadoop.io.LongWritable;
 import  org.apache.hadoop.util.StringUtils;
 
Compute the variance. This class is extended by: GenericUDAFVarianceSample GenericUDAFStd GenericUDAFStdSample
 
 @Description(name = "variance,var_pop", value = "_FUNC_(x) - Returns the variance of a set of numbers")
 public class GenericUDAFVariance extends AbstractGenericUDAFResolver {
 
     static final Log LOG = LogFactory.getLog(GenericUDAFVariance.class.getName());
 
     @Override
     public GenericUDAFEvaluator getEvaluator(TypeInfo[] parametersthrows SemanticException {
         if (parameters.length != 1) {
             throw new UDFArgumentTypeException(parameters.length - 1, "Exactly one argument is expected.");
         }
 
         if (parameters[0].getCategory() != ObjectInspector.Category.PRIMITIVE) {
             throw new UDFArgumentTypeException(0, "Only primitive type arguments are accepted but "
                     + parameters[0].getTypeName() + " is passed.");
         }
         switch (((PrimitiveTypeInfo) parameters[0]).getPrimitiveCategory()) {
             case BYTE:
             case SHORT:
             case INT:
             case LONG:
             case FLOAT:
             case DOUBLE:
             case STRING:
                 return new GenericUDAFVarianceEvaluator();
             case BOOLEAN:
             default:
                 throw new UDFArgumentTypeException(0, "Only numeric or string type arguments are accepted but "
                         + parameters[0].getTypeName() + " is passed.");
         }
     }

    
Evaluate the variance using the algorithm described by Chan, Golub, and LeVeque in "Algorithms for computing the sample variance: analysis and recommendations" The American Statistician, 37 (1983) pp. 242--247. variance = variance1 + variance2 + n/(m*(m+n)) * pow(((m/n)*t1 - t2),2) where: - variance is sum[x-avg^2] (this is actually n times the variance) and is updated at every step. - n is the count of elements in chunk1 - m is the count of elements in chunk2 - t1 = sum of elements in chunk1, t2 = sum of elements in chunk2. This algorithm was proven to be numerically stable by J.L. Barlow in "Error analysis of a pairwise summation algorithm to compute sample variance" Numer. Math, 58 (1991) pp. 583--590
 
     public static class GenericUDAFVarianceEvaluator extends GenericUDAFEvaluator {
 
         // For PARTIAL1 and COMPLETE
        private PrimitiveObjectInspector inputOI;
        // For PARTIAL2 and FINAL
        private StructObjectInspector soi;
        private StructField countField;
        private StructField sumField;
        private StructField varianceField;
        private LongObjectInspector countFieldOI;
        private DoubleObjectInspector sumFieldOI;
        // For PARTIAL1 and PARTIAL2
        private Object[] partialResult;
        // For FINAL and COMPLETE
        private DoubleWritable result;
        @Override
        public ObjectInspector init(Mode m, ObjectInspector[] parametersthrows HiveException {
            assert (parameters.length == 1);
            super.init(mparameters);
            // init input
            if (mode == Mode.PARTIAL1 || mode == Mode.COMPLETE) {
                 = (PrimitiveObjectInspector) parameters[0];
            } else {
                 = (StructObjectInspector) parameters[0];
                 = .getStructFieldRef("count");
                 = .getStructFieldRef("sum");
                 = .getStructFieldRef("variance");
                 = (LongObjectInspector) .getFieldObjectInspector();
                 = (DoubleObjectInspector) .getFieldObjectInspector();
            }
            // init output
            if (mode == Mode.PARTIAL1 || mode == Mode.PARTIAL2) {
                // The output of a partial aggregation is a struct containing
                // a long count and doubles sum and variance.
                ArrayList<ObjectInspector> foi = new ArrayList<ObjectInspector>();
                foi.add(PrimitiveObjectInspectorFactory.writableLongObjectInspector);
                foi.add(PrimitiveObjectInspectorFactory.writableDoubleObjectInspector);
                foi.add(PrimitiveObjectInspectorFactory.writableDoubleObjectInspector);
                ArrayList<Stringfname = new ArrayList<String>();
                fname.add("count");
                fname.add("sum");
                fname.add("variance");
                 = new Object[3];
                [0] = new LongWritable(0);
                [1] = new DoubleWritable(0);
                [2] = new DoubleWritable(0);
                return ObjectInspectorFactory.getStandardStructObjectInspector(fnamefoi);
            } else {
                setResult(new DoubleWritable(0));
                return PrimitiveObjectInspectorFactory.writableDoubleObjectInspector;
            }
        }
        static class StdAgg implements SerializableBuffer {
            long count// number of elements
            double sum// sum of elements
            double variance// sum[x-avg^2] (this is actually n times the
                             // variance)
            @Override
            public void deSerializeAggBuffer(byte[] dataint startint len) {
                 = BufferSerDeUtil.getLong(datastart);
                start += 8;
                 = BufferSerDeUtil.getDouble(datastart);
                start += 8;
                 = BufferSerDeUtil.getDouble(datastart);
            }
            @Override
            public void serializeAggBuffer(byte[] dataint startint len) {
                BufferSerDeUtil.writeLong(datastart);
                start += 8;
                BufferSerDeUtil.writeDouble(datastart);
                start += 8;
                BufferSerDeUtil.writeDouble(datastart);
            }
            @Override
            public void serializeAggBuffer(DataOutput outputthrows IOException {
                output.writeLong();
                output.writeDouble();
                output.writeDouble();
            }
        };
        @Override
        public AggregationBuffer getNewAggregationBuffer() throws HiveException {
            StdAgg result = new StdAgg();
            reset(result);
            return result;
        }
        @Override
        public void reset(AggregationBuffer aggthrows HiveException {
            StdAgg myagg = (StdAggagg;
            myagg.count = 0;
            myagg.sum = 0;
            myagg.variance = 0;
        }
        private boolean warned = false;
        @Override
        public void iterate(AggregationBuffer aggObject[] parametersthrows HiveException {
            assert (parameters.length == 1);
            Object p = parameters[0];
            if (p != null) {
                StdAgg myagg = (StdAggagg;
                try {
                    double v = PrimitiveObjectInspectorUtils.getDouble(p);
                    myagg.count++;
                    myagg.sum += v;
                    if (myagg.count > 1) {
                        double t = myagg.count * v - myagg.sum;
                        myagg.variance += (t * t) / ((doublemyagg.count * (myagg.count - 1));
                    }
                } catch (NumberFormatException e) {
                    if (!) {
                         = true;
                        .warn(getClass().getSimpleName() + " " + StringUtils.stringifyException(e));
                        .warn(getClass().getSimpleName() + " ignoring similar exceptions.");
                    }
                }
            }
        }
        @Override
        public Object terminatePartial(AggregationBuffer aggthrows HiveException {
            StdAgg myagg = (StdAggagg;
            ((LongWritable) [0]).set(myagg.count);
            ((DoubleWritable) [1]).set(myagg.sum);
            ((DoubleWritable) [2]).set(myagg.variance);
            return ;
        }
        @Override
        public void merge(AggregationBuffer aggObject partialthrows HiveException {
            if (partial != null) {
                StdAgg myagg = (StdAggagg;
                Object partialCount = .getStructFieldData(partial);
                Object partialSum = .getStructFieldData(partial);
                Object partialVariance = .getStructFieldData(partial);
                long n = myagg.count;
                long m = .get(partialCount);
                if (n == 0) {
                    // Just copy the information since there is nothing so far
                    myagg.variance = .get(partialVariance);
                    myagg.count = .get(partialCount);
                    myagg.sum = .get(partialSum);
                }
                if (m != 0 && n != 0) {
                    // Merge the two partials
                    double a = myagg.sum;
                    double b = .get(partialSum);
                    myagg.count += m;
                    myagg.sum += b;
                    double t = (m / (doublen) * a - b;
                    myagg.variance += .get(partialVariance) + ((n / (doublem) / ((doublen + m)) * t * t;
                }
            }
        }
        @Override
        public Object terminate(AggregationBuffer aggthrows HiveException {
            StdAgg myagg = (StdAggagg;
            if (myagg.count == 0) { // SQL standard - return null for zero
                                    // elements
                return null;
            } else {
                if (myagg.count > 1) {
                    getResult().set(myagg.variance / (myagg.count));
                } else { // for one element the variance is always 0
                    getResult().set(0);
                }
                return getResult();
            }
        }
        public void setResult(DoubleWritable result) {
            this. = result;
        }
        public DoubleWritable getResult() {
            return ;
        }
    }
New to GrepCode? Check out our FAQ X