Search in sources :

Example 1 with ListObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector in project hive by apache.

the class WindowingTableFunction method iterator.

@SuppressWarnings("rawtypes")
@Override
public Iterator<Object> iterator(PTFPartitionIterator<Object> pItr) throws HiveException {
    WindowTableFunctionDef wTFnDef = (WindowTableFunctionDef) getTableDef();
    ArrayList<Object> output = new ArrayList<Object>();
    List<?>[] outputFromPivotFunctions = new List<?>[wTFnDef.getWindowFunctions().size()];
    ArrayList<Integer> wFnsWithWindows = new ArrayList<Integer>();
    PTFPartition iPart = pItr.getPartition();
    int i = 0;
    for (WindowFunctionDef wFn : wTFnDef.getWindowFunctions()) {
        boolean processWindow = processWindow(wFn.getWindowFrame());
        pItr.reset();
        if (!processWindow && !wFn.isPivotResult()) {
            Object out = evaluateFunctionOnPartition(wFn, iPart);
            output.add(out);
        } else if (wFn.isPivotResult()) {
            GenericUDAFEvaluator streamingEval = wFn.getWFnEval().getWindowingEvaluator(wFn.getWindowFrame());
            if (streamingEval != null && streamingEval instanceof ISupportStreamingModeForWindowing) {
                ISupportStreamingModeForWindowing strEval = (ISupportStreamingModeForWindowing) streamingEval;
                if (strEval.getRowsRemainingAfterTerminate() == 0) {
                    wFn.setWFnEval(streamingEval);
                    if (wFn.getOI() instanceof ListObjectInspector) {
                        ListObjectInspector listOI = (ListObjectInspector) wFn.getOI();
                        wFn.setOI(listOI.getListElementObjectInspector());
                    }
                    output.add(null);
                    wFnsWithWindows.add(i);
                } else {
                    outputFromPivotFunctions[i] = (List) evaluateFunctionOnPartition(wFn, iPart);
                    output.add(null);
                }
            } else {
                outputFromPivotFunctions[i] = (List) evaluateFunctionOnPartition(wFn, iPart);
                output.add(null);
            }
        } else {
            output.add(null);
            wFnsWithWindows.add(i);
        }
        i++;
    }
    for (i = 0; i < iPart.getOutputOI().getAllStructFieldRefs().size(); i++) {
        output.add(null);
    }
    if (wTFnDef.getRankLimit() != -1) {
        rnkLimitDef = new RankLimit(wTFnDef.getRankLimit(), wTFnDef.getRankLimitFunction(), wTFnDef.getWindowFunctions());
    }
    return new WindowingIterator(iPart, output, outputFromPivotFunctions, ArrayUtils.toPrimitive(wFnsWithWindows.toArray(new Integer[wFnsWithWindows.size()])));
}
Also used : ISupportStreamingModeForWindowing(org.apache.hadoop.hive.ql.udf.generic.ISupportStreamingModeForWindowing) GenericUDAFEvaluator(org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator) PTFPartition(org.apache.hadoop.hive.ql.exec.PTFPartition) WindowTableFunctionDef(org.apache.hadoop.hive.ql.plan.ptf.WindowTableFunctionDef) ArrayList(java.util.ArrayList) ListObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector) AbstractList(java.util.AbstractList) ArrayList(java.util.ArrayList) List(java.util.List) WindowFunctionDef(org.apache.hadoop.hive.ql.plan.ptf.WindowFunctionDef)

Example 2 with ListObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector in project hive by apache.

the class ObjectInspectorConverters method getConvertedOI.

/**
   * Utility function to convert from one object inspector type to another.
   * The output object inspector type should have all fields as settableOI type.
   * The above condition can be violated only if equalsCheck is true and inputOI is
   * equal to outputOI.
   * @param inputOI : input object inspector
   * @param outputOI : output object inspector
   * @param oiSettableProperties : The object inspector to isSettable mapping used to cache
   *                               intermediate results.
   * @param equalsCheck : Do we need to check if the inputOI and outputOI are the same?
   *                      true : If they are the same, we return the object inspector directly.
   *                      false : Do not perform an equality check on inputOI and outputOI
   * @return : The output object inspector containing all settable fields. The return value
   *           can contain non-settable fields only if inputOI equals outputOI and equalsCheck is
   *           true.
   */
public static ObjectInspector getConvertedOI(ObjectInspector inputOI, ObjectInspector outputOI, Map<ObjectInspector, Boolean> oiSettableProperties, boolean equalsCheck) {
    // 2. If the outputOI has all fields settable, return it
    if ((equalsCheck && inputOI.equals(outputOI)) || ObjectInspectorUtils.hasAllFieldsSettable(outputOI, oiSettableProperties) == true) {
        return outputOI;
    }
    // T is settable recursively i.e all the nested fields are also settable.
    switch(outputOI.getCategory()) {
        case PRIMITIVE:
            // Create a writable object inspector for primitive type and return it.
            PrimitiveObjectInspector primOutputOI = (PrimitiveObjectInspector) outputOI;
            return PrimitiveObjectInspectorFactory.getPrimitiveWritableObjectInspector(primOutputOI.getTypeInfo());
        case STRUCT:
            StructObjectInspector structOutputOI = (StructObjectInspector) outputOI;
            // create a standard settable struct object inspector.
            List<? extends StructField> listFields = structOutputOI.getAllStructFieldRefs();
            List<String> structFieldNames = new ArrayList<String>(listFields.size());
            List<ObjectInspector> structFieldObjectInspectors = new ArrayList<ObjectInspector>(listFields.size());
            for (StructField listField : listFields) {
                structFieldNames.add(listField.getFieldName());
                // We need to make sure that the underlying fields are settable as well.
                // Hence, the recursive call for each field.
                // Note that equalsCheck is false while invoking getConvertedOI() because
                // we need to bypass the initial inputOI.equals(outputOI) check.
                structFieldObjectInspectors.add(getConvertedOI(listField.getFieldObjectInspector(), listField.getFieldObjectInspector(), oiSettableProperties, false));
            }
            return ObjectInspectorFactory.getStandardStructObjectInspector(structFieldNames, structFieldObjectInspectors);
        case LIST:
            ListObjectInspector listOutputOI = (ListObjectInspector) outputOI;
            // We need to make sure that the list element type is settable.
            return ObjectInspectorFactory.getStandardListObjectInspector(getConvertedOI(listOutputOI.getListElementObjectInspector(), listOutputOI.getListElementObjectInspector(), oiSettableProperties, false));
        case MAP:
            MapObjectInspector mapOutputOI = (MapObjectInspector) outputOI;
            // We need to make sure that the key type and the value types are settable.
            return ObjectInspectorFactory.getStandardMapObjectInspector(getConvertedOI(mapOutputOI.getMapKeyObjectInspector(), mapOutputOI.getMapKeyObjectInspector(), oiSettableProperties, false), getConvertedOI(mapOutputOI.getMapValueObjectInspector(), mapOutputOI.getMapValueObjectInspector(), oiSettableProperties, false));
        case UNION:
            UnionObjectInspector unionOutputOI = (UnionObjectInspector) outputOI;
            // create a standard settable union object inspector
            List<ObjectInspector> unionListFields = unionOutputOI.getObjectInspectors();
            List<ObjectInspector> unionFieldObjectInspectors = new ArrayList<ObjectInspector>(unionListFields.size());
            for (ObjectInspector listField : unionListFields) {
                // We need to make sure that all the field associated with the union are settable.
                unionFieldObjectInspectors.add(getConvertedOI(listField, listField, oiSettableProperties, false));
            }
            return ObjectInspectorFactory.getStandardUnionObjectInspector(unionFieldObjectInspectors);
        default:
            // Unsupported in-memory structure.
            throw new RuntimeException("Hive internal error: conversion of " + inputOI.getTypeName() + " to " + outputOI.getTypeName() + " not supported yet.");
    }
}
Also used : SettableDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDoubleObjectInspector) VoidObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.VoidObjectInspector) SettableBooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBooleanObjectInspector) SettableHiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveCharObjectInspector) SettableShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableShortObjectInspector) SettableTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableTimestampObjectInspector) SettableHiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalDayTimeObjectInspector) SettableLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableLongObjectInspector) SettableBinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBinaryObjectInspector) SettableFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableFloatObjectInspector) SettableHiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveDecimalObjectInspector) JavaStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.JavaStringObjectInspector) SettableHiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalYearMonthObjectInspector) SettableHiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveVarcharObjectInspector) SettableIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableIntObjectInspector) SettableDateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDateObjectInspector) WritableStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.WritableStringObjectInspector) SettableByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableByteObjectInspector) ArrayList(java.util.ArrayList)

Example 3 with ListObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector in project hive by apache.

the class ObjectInspectorUtils method hasAllFieldsSettable.

/**
   *
   * @param oi - Input object inspector
   * @param oiSettableProperties - Lookup map to cache the result.(If no caching, pass null)
   * @return - true if : (1) oi is an instance of settable<DataType>OI.
   *                     (2) All the embedded object inspectors are instances of settable<DataType>OI.
   *           If (1) or (2) is false, return false.
   */
public static boolean hasAllFieldsSettable(ObjectInspector oi, Map<ObjectInspector, Boolean> oiSettableProperties) {
    // If the result is already present in the cache, return it.
    if (!(oiSettableProperties == null) && oiSettableProperties.containsKey(oi)) {
        return oiSettableProperties.get(oi).booleanValue();
    }
    // If the top-level object inspector is non-settable return false
    if (!(isInstanceOfSettableOI(oi))) {
        return setOISettablePropertiesMap(oi, oiSettableProperties, false);
    }
    Boolean returnValue = true;
    switch(oi.getCategory()) {
        case PRIMITIVE:
            break;
        case STRUCT:
            StructObjectInspector structOutputOI = (StructObjectInspector) oi;
            List<? extends StructField> listFields = structOutputOI.getAllStructFieldRefs();
            for (StructField listField : listFields) {
                if (!hasAllFieldsSettable(listField.getFieldObjectInspector(), oiSettableProperties)) {
                    returnValue = false;
                    break;
                }
            }
            break;
        case LIST:
            ListObjectInspector listOutputOI = (ListObjectInspector) oi;
            returnValue = hasAllFieldsSettable(listOutputOI.getListElementObjectInspector(), oiSettableProperties);
            break;
        case MAP:
            MapObjectInspector mapOutputOI = (MapObjectInspector) oi;
            returnValue = hasAllFieldsSettable(mapOutputOI.getMapKeyObjectInspector(), oiSettableProperties) && hasAllFieldsSettable(mapOutputOI.getMapValueObjectInspector(), oiSettableProperties);
            break;
        case UNION:
            UnionObjectInspector unionOutputOI = (UnionObjectInspector) oi;
            List<ObjectInspector> unionListFields = unionOutputOI.getObjectInspectors();
            for (ObjectInspector listField : unionListFields) {
                if (!hasAllFieldsSettable(listField, oiSettableProperties)) {
                    returnValue = false;
                    break;
                }
            }
            break;
        default:
            throw new RuntimeException("Hive internal error inside hasAllFieldsSettable : " + oi.getTypeName() + " not supported yet.");
    }
    return setOISettablePropertiesMap(oi, oiSettableProperties, returnValue);
}
Also used : SettableTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableTimestampObjectInspector) SettableBinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBinaryObjectInspector) SettableHiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveDecimalObjectInspector) HiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveDecimalObjectInspector) BooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.BooleanObjectInspector) ShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector) HiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveIntervalYearMonthObjectInspector) SettableHiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveCharObjectInspector) SettableLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableLongObjectInspector) FloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.FloatObjectInspector) StringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.StringObjectInspector) HiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveIntervalDayTimeObjectInspector) DateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DateObjectInspector) SettableDateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDateObjectInspector) WritableStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.WritableStringObjectInspector) HiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveVarcharObjectInspector) SettableShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableShortObjectInspector) SettableHiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalDayTimeObjectInspector) HiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveCharObjectInspector) IntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.IntObjectInspector) SettableFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableFloatObjectInspector) AbstractPrimitiveWritableObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.AbstractPrimitiveWritableObjectInspector) SettableHiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveVarcharObjectInspector) LongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector) BinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.BinaryObjectInspector) SettableDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDoubleObjectInspector) SettableBooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBooleanObjectInspector) JavaStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.JavaStringObjectInspector) SettableHiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalYearMonthObjectInspector) SettableIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableIntObjectInspector) ByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.ByteObjectInspector) DoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector) TimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.TimestampObjectInspector) SettableByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableByteObjectInspector)

Example 4 with ListObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector in project hive by apache.

the class ObjectInspectorUtils method getConstantObjectInspector.

public static ConstantObjectInspector getConstantObjectInspector(ObjectInspector oi, Object value) {
    if (oi instanceof ConstantObjectInspector) {
        return (ConstantObjectInspector) oi;
    }
    ObjectInspector writableOI = getStandardObjectInspector(oi, ObjectInspectorCopyOption.WRITABLE);
    Object writableValue = value == null ? value : ObjectInspectorConverters.getConverter(oi, writableOI).convert(value);
    switch(writableOI.getCategory()) {
        case PRIMITIVE:
            PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi;
            return PrimitiveObjectInspectorFactory.getPrimitiveWritableConstantObjectInspector(poi.getTypeInfo(), writableValue);
        case LIST:
            ListObjectInspector loi = (ListObjectInspector) oi;
            return ObjectInspectorFactory.getStandardConstantListObjectInspector(getStandardObjectInspector(loi.getListElementObjectInspector(), ObjectInspectorCopyOption.WRITABLE), (List<?>) writableValue);
        case MAP:
            MapObjectInspector moi = (MapObjectInspector) oi;
            return ObjectInspectorFactory.getStandardConstantMapObjectInspector(getStandardObjectInspector(moi.getMapKeyObjectInspector(), ObjectInspectorCopyOption.WRITABLE), getStandardObjectInspector(moi.getMapValueObjectInspector(), ObjectInspectorCopyOption.WRITABLE), (Map<?, ?>) writableValue);
        case STRUCT:
            StructObjectInspector soi = (StructObjectInspector) oi;
            List<? extends StructField> fields = soi.getAllStructFieldRefs();
            List<String> fieldNames = new ArrayList<String>(fields.size());
            List<ObjectInspector> fieldObjectInspectors = new ArrayList<ObjectInspector>(fields.size());
            for (StructField f : fields) {
                fieldNames.add(f.getFieldName());
                fieldObjectInspectors.add(getStandardObjectInspector(f.getFieldObjectInspector(), ObjectInspectorCopyOption.WRITABLE));
            }
            if (value != null && (writableValue.getClass().isArray())) {
                writableValue = java.util.Arrays.asList((Object[]) writableValue);
            }
            return ObjectInspectorFactory.getStandardConstantStructObjectInspector(fieldNames, fieldObjectInspectors, (List<?>) writableValue);
        default:
            throw new IllegalArgumentException(writableOI.getCategory() + " not yet supported for constant OI");
    }
}
Also used : SettableTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableTimestampObjectInspector) SettableBinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBinaryObjectInspector) SettableHiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveDecimalObjectInspector) HiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveDecimalObjectInspector) BooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.BooleanObjectInspector) ShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector) HiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveIntervalYearMonthObjectInspector) SettableHiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveCharObjectInspector) SettableLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableLongObjectInspector) FloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.FloatObjectInspector) StringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.StringObjectInspector) HiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveIntervalDayTimeObjectInspector) DateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DateObjectInspector) SettableDateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDateObjectInspector) WritableStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.WritableStringObjectInspector) HiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveVarcharObjectInspector) SettableShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableShortObjectInspector) SettableHiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalDayTimeObjectInspector) HiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveCharObjectInspector) IntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.IntObjectInspector) SettableFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableFloatObjectInspector) AbstractPrimitiveWritableObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.AbstractPrimitiveWritableObjectInspector) SettableHiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveVarcharObjectInspector) LongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector) BinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.BinaryObjectInspector) SettableDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDoubleObjectInspector) SettableBooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBooleanObjectInspector) JavaStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.JavaStringObjectInspector) SettableHiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalYearMonthObjectInspector) SettableIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableIntObjectInspector) ByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.ByteObjectInspector) DoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector) TimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.TimestampObjectInspector) SettableByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableByteObjectInspector) ArrayList(java.util.ArrayList)

Example 5 with ListObjectInspector

use of org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector in project hive by apache.

the class ObjectInspectorUtils method compare.

/**
   * Compare two objects with their respective ObjectInspectors.
   * if nullValueOpt is MAXVALUE, treat null as maximum value.
   * if nullValueOpt is MINVALUE, treat null as minimum value.
   */
public static int compare(Object o1, ObjectInspector oi1, Object o2, ObjectInspector oi2, MapEqualComparer mapEqualComparer, NullValueOption nullValueOpt) {
    if (oi1.getCategory() != oi2.getCategory()) {
        return oi1.getCategory().compareTo(oi2.getCategory());
    }
    int nullCmpRtn = -1;
    switch(nullValueOpt) {
        case MAXVALUE:
            nullCmpRtn = 1;
            break;
        case MINVALUE:
            nullCmpRtn = -1;
            break;
    }
    if (o1 == null) {
        return o2 == null ? 0 : nullCmpRtn;
    } else if (o2 == null) {
        return -nullCmpRtn;
    }
    switch(oi1.getCategory()) {
        case PRIMITIVE:
            {
                PrimitiveObjectInspector poi1 = ((PrimitiveObjectInspector) oi1);
                PrimitiveObjectInspector poi2 = ((PrimitiveObjectInspector) oi2);
                if (poi1.getPrimitiveCategory() != poi2.getPrimitiveCategory()) {
                    return poi1.getPrimitiveCategory().compareTo(poi2.getPrimitiveCategory());
                }
                switch(poi1.getPrimitiveCategory()) {
                    case VOID:
                        return 0;
                    case BOOLEAN:
                        {
                            int v1 = ((BooleanObjectInspector) poi1).get(o1) ? 1 : 0;
                            int v2 = ((BooleanObjectInspector) poi2).get(o2) ? 1 : 0;
                            return v1 - v2;
                        }
                    case BYTE:
                        {
                            int v1 = ((ByteObjectInspector) poi1).get(o1);
                            int v2 = ((ByteObjectInspector) poi2).get(o2);
                            return v1 - v2;
                        }
                    case SHORT:
                        {
                            int v1 = ((ShortObjectInspector) poi1).get(o1);
                            int v2 = ((ShortObjectInspector) poi2).get(o2);
                            return v1 - v2;
                        }
                    case INT:
                        {
                            int v1 = ((IntObjectInspector) poi1).get(o1);
                            int v2 = ((IntObjectInspector) poi2).get(o2);
                            return v1 > v2 ? 1 : (v1 < v2 ? -1 : 0);
                        }
                    case LONG:
                        {
                            long v1 = ((LongObjectInspector) poi1).get(o1);
                            long v2 = ((LongObjectInspector) poi2).get(o2);
                            return v1 > v2 ? 1 : (v1 < v2 ? -1 : 0);
                        }
                    case FLOAT:
                        {
                            float v1 = ((FloatObjectInspector) poi1).get(o1);
                            float v2 = ((FloatObjectInspector) poi2).get(o2);
                            // The IEEE 754 floating point spec specifies that signed -0.0 and 0.0 should be treated as equal.
                            if (v1 == 0.0f && v2 == 0.0f) {
                                return 0;
                            } else {
                                // Float.compare() treats -0.0 and 0.0 as different
                                return Float.compare(v1, v2);
                            }
                        }
                    case DOUBLE:
                        {
                            double v1 = ((DoubleObjectInspector) poi1).get(o1);
                            double v2 = ((DoubleObjectInspector) poi2).get(o2);
                            // The IEEE 754 floating point spec specifies that signed -0.0 and 0.0 should be treated as equal.
                            if (v1 == 0.0d && v2 == 0.0d) {
                                return 0;
                            } else {
                                // Double.compare() treats -0.0 and 0.0 as different
                                return Double.compare(v1, v2);
                            }
                        }
                    case STRING:
                        {
                            if (poi1.preferWritable() || poi2.preferWritable()) {
                                Text t1 = (Text) poi1.getPrimitiveWritableObject(o1);
                                Text t2 = (Text) poi2.getPrimitiveWritableObject(o2);
                                return t1 == null ? (t2 == null ? 0 : -1) : (t2 == null ? 1 : t1.compareTo(t2));
                            } else {
                                String s1 = (String) poi1.getPrimitiveJavaObject(o1);
                                String s2 = (String) poi2.getPrimitiveJavaObject(o2);
                                return s1 == null ? (s2 == null ? 0 : -1) : (s2 == null ? 1 : s1.compareTo(s2));
                            }
                        }
                    case CHAR:
                        {
                            HiveCharWritable t1 = ((HiveCharObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            HiveCharWritable t2 = ((HiveCharObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return t1.compareTo(t2);
                        }
                    case VARCHAR:
                        {
                            HiveVarcharWritable t1 = ((HiveVarcharObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            HiveVarcharWritable t2 = ((HiveVarcharObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return t1.compareTo(t2);
                        }
                    case BINARY:
                        {
                            BytesWritable bw1 = ((BinaryObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            BytesWritable bw2 = ((BinaryObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return bw1.compareTo(bw2);
                        }
                    case DATE:
                        {
                            DateWritable d1 = ((DateObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            DateWritable d2 = ((DateObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return d1.compareTo(d2);
                        }
                    case TIMESTAMP:
                        {
                            TimestampWritable t1 = ((TimestampObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            TimestampWritable t2 = ((TimestampObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return t1.compareTo(t2);
                        }
                    case INTERVAL_YEAR_MONTH:
                        {
                            HiveIntervalYearMonthWritable i1 = ((HiveIntervalYearMonthObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            HiveIntervalYearMonthWritable i2 = ((HiveIntervalYearMonthObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return i1.compareTo(i2);
                        }
                    case INTERVAL_DAY_TIME:
                        {
                            HiveIntervalDayTimeWritable i1 = ((HiveIntervalDayTimeObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            HiveIntervalDayTimeWritable i2 = ((HiveIntervalDayTimeObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return i1.compareTo(i2);
                        }
                    case DECIMAL:
                        {
                            HiveDecimalWritable t1 = ((HiveDecimalObjectInspector) poi1).getPrimitiveWritableObject(o1);
                            HiveDecimalWritable t2 = ((HiveDecimalObjectInspector) poi2).getPrimitiveWritableObject(o2);
                            return t1.compareTo(t2);
                        }
                    default:
                        {
                            throw new RuntimeException("Unknown type: " + poi1.getPrimitiveCategory());
                        }
                }
            }
        case STRUCT:
            {
                StructObjectInspector soi1 = (StructObjectInspector) oi1;
                StructObjectInspector soi2 = (StructObjectInspector) oi2;
                List<? extends StructField> fields1 = soi1.getAllStructFieldRefs();
                List<? extends StructField> fields2 = soi2.getAllStructFieldRefs();
                int minimum = Math.min(fields1.size(), fields2.size());
                for (int i = 0; i < minimum; i++) {
                    int r = compare(soi1.getStructFieldData(o1, fields1.get(i)), fields1.get(i).getFieldObjectInspector(), soi2.getStructFieldData(o2, fields2.get(i)), fields2.get(i).getFieldObjectInspector(), mapEqualComparer, nullValueOpt);
                    if (r != 0) {
                        return r;
                    }
                }
                return fields1.size() - fields2.size();
            }
        case LIST:
            {
                ListObjectInspector loi1 = (ListObjectInspector) oi1;
                ListObjectInspector loi2 = (ListObjectInspector) oi2;
                int minimum = Math.min(loi1.getListLength(o1), loi2.getListLength(o2));
                for (int i = 0; i < minimum; i++) {
                    int r = compare(loi1.getListElement(o1, i), loi1.getListElementObjectInspector(), loi2.getListElement(o2, i), loi2.getListElementObjectInspector(), mapEqualComparer, nullValueOpt);
                    if (r != 0) {
                        return r;
                    }
                }
                return loi1.getListLength(o1) - loi2.getListLength(o2);
            }
        case MAP:
            {
                if (mapEqualComparer == null) {
                    throw new RuntimeException("Compare on map type not supported!");
                } else {
                    return mapEqualComparer.compare(o1, (MapObjectInspector) oi1, o2, (MapObjectInspector) oi2);
                }
            }
        case UNION:
            {
                UnionObjectInspector uoi1 = (UnionObjectInspector) oi1;
                UnionObjectInspector uoi2 = (UnionObjectInspector) oi2;
                byte tag1 = uoi1.getTag(o1);
                byte tag2 = uoi2.getTag(o2);
                if (tag1 != tag2) {
                    return tag1 - tag2;
                }
                return compare(uoi1.getField(o1), uoi1.getObjectInspectors().get(tag1), uoi2.getField(o2), uoi2.getObjectInspectors().get(tag2), mapEqualComparer, nullValueOpt);
            }
        default:
            throw new RuntimeException("Compare on unknown type: " + oi1.getCategory());
    }
}
Also used : SettableLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableLongObjectInspector) LongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector) DateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DateObjectInspector) SettableDateObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDateObjectInspector) IntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.IntObjectInspector) SettableIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableIntObjectInspector) SettableBinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableBinaryObjectInspector) BinaryObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.BinaryObjectInspector) TimestampWritable(org.apache.hadoop.hive.serde2.io.TimestampWritable) FloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.FloatObjectInspector) SettableFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableFloatObjectInspector) ByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.ByteObjectInspector) SettableByteObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableByteObjectInspector) SettableTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableTimestampObjectInspector) TimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.TimestampObjectInspector) ShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector) SettableShortObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableShortObjectInspector) SettableHiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveDecimalObjectInspector) HiveDecimalObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveDecimalObjectInspector) List(java.util.List) ArrayList(java.util.ArrayList) SettableHiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveCharObjectInspector) HiveCharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveCharObjectInspector) DateWritable(org.apache.hadoop.hive.serde2.io.DateWritable) HiveDecimalWritable(org.apache.hadoop.hive.serde2.io.HiveDecimalWritable) HiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveIntervalDayTimeObjectInspector) SettableHiveIntervalDayTimeObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalDayTimeObjectInspector) HiveCharWritable(org.apache.hadoop.hive.serde2.io.HiveCharWritable) HiveVarcharWritable(org.apache.hadoop.hive.serde2.io.HiveVarcharWritable) Text(org.apache.hadoop.io.Text) BytesWritable(org.apache.hadoop.io.BytesWritable) HiveIntervalDayTimeWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalDayTimeWritable) HiveIntervalYearMonthWritable(org.apache.hadoop.hive.serde2.io.HiveIntervalYearMonthWritable) SettableDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableDoubleObjectInspector) DoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector) HiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveVarcharObjectInspector) SettableHiveVarcharObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveVarcharObjectInspector) HiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveIntervalYearMonthObjectInspector) SettableHiveIntervalYearMonthObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.SettableHiveIntervalYearMonthObjectInspector)

Aggregations

ListObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector)48 ObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector)35 StructObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector)34 MapObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.MapObjectInspector)26 PrimitiveObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector)25 ArrayList (java.util.ArrayList)23 LongObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.LongObjectInspector)17 TimestampObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.TimestampObjectInspector)17 DoubleObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.DoubleObjectInspector)16 IntObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.IntObjectInspector)16 ShortObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.ShortObjectInspector)16 ByteObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.ByteObjectInspector)15 FloatObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.FloatObjectInspector)15 StringObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.StringObjectInspector)15 BooleanObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.BooleanObjectInspector)14 HiveDecimalObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.HiveDecimalObjectInspector)14 List (java.util.List)13 BinaryObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.BinaryObjectInspector)13 DateObjectInspector (org.apache.hadoop.hive.serde2.objectinspector.primitive.DateObjectInspector)13 Map (java.util.Map)11