Examples with FilterNumericDocValues org.apache.lucene.index.FilterNumericDocValues used on opensource projects

Example 1 with FilterNumericDocValues

use of org.apache.lucene.index.FilterNumericDocValues in project lucene-solr by apache.

the class IntervalFacets method getCountNumeric.

private void getCountNumeric() throws IOException {
    final FieldType ft = schemaField.getType();
    final String fieldName = schemaField.getName();
    final NumberType numericType = ft.getNumberType();
    if (numericType == null) {
        throw new IllegalStateException();
    }
    final List<LeafReaderContext> leaves = searcher.getIndexReader().leaves();
    final Iterator<LeafReaderContext> ctxIt = leaves.iterator();
    LeafReaderContext ctx = null;
    NumericDocValues longs = null;
    for (DocIterator docsIt = docs.iterator(); docsIt.hasNext(); ) {
        final int doc = docsIt.nextDoc();
        if (ctx == null || doc >= ctx.docBase + ctx.reader().maxDoc()) {
            do {
                ctx = ctxIt.next();
            } while (ctx == null || doc >= ctx.docBase + ctx.reader().maxDoc());
            assert doc >= ctx.docBase;
            switch(numericType) {
                case LONG:
                case DATE:
                case INTEGER:
                    longs = DocValues.getNumeric(ctx.reader(), fieldName);
                    break;
                case FLOAT:
                    // TODO: this bit flipping should probably be moved to tie-break in the PQ comparator
                    longs = new FilterNumericDocValues(DocValues.getNumeric(ctx.reader(), fieldName)) {

                        @Override
                        public long longValue() throws IOException {
                            long bits = super.longValue();
                            if (bits < 0)
                                bits ^= 0x7fffffffffffffffL;
                            return bits;
                        }
                    };
                    break;
                case DOUBLE:
                    // TODO: this bit flipping should probably be moved to tie-break in the PQ comparator
                    longs = new FilterNumericDocValues(DocValues.getNumeric(ctx.reader(), fieldName)) {

                        @Override
                        public long longValue() throws IOException {
                            long bits = super.longValue();
                            if (bits < 0)
                                bits ^= 0x7fffffffffffffffL;
                            return bits;
                        }
                    };
                    break;
                default:
                    throw new AssertionError();
            }
        }
        int valuesDocID = longs.docID();
        if (valuesDocID < doc - ctx.docBase) {
            valuesDocID = longs.advance(doc - ctx.docBase);
        }
        if (valuesDocID == doc - ctx.docBase) {
            accumIntervalWithValue(longs.longValue());
        }
    }
}

Example 2 with FilterNumericDocValues

use of org.apache.lucene.index.FilterNumericDocValues in project lucene-solr by apache.

the class NumericFacets method getCountsSingleValue.

private static NamedList<Integer> getCountsSingleValue(SolrIndexSearcher searcher, DocSet docs, String fieldName, int offset, int limit, int mincount, boolean missing, String sort) throws IOException {
    boolean zeros = mincount <= 0;
    mincount = Math.max(mincount, 1);
    final SchemaField sf = searcher.getSchema().getField(fieldName);
    final FieldType ft = sf.getType();
    final NumberType numericType = ft.getNumberType();
    if (numericType == null) {
        throw new IllegalStateException();
    }
    // We don't return zeros when using PointFields or when index=false
    zeros = zeros && !ft.isPointField() && sf.indexed();
    final List<LeafReaderContext> leaves = searcher.getIndexReader().leaves();
    // 1. accumulate
    final HashTable hashTable = new HashTable(true);
    final Iterator<LeafReaderContext> ctxIt = leaves.iterator();
    LeafReaderContext ctx = null;
    NumericDocValues longs = null;
    int missingCount = 0;
    for (DocIterator docsIt = docs.iterator(); docsIt.hasNext(); ) {
        final int doc = docsIt.nextDoc();
        if (ctx == null || doc >= ctx.docBase + ctx.reader().maxDoc()) {
            do {
                ctx = ctxIt.next();
            } while (ctx == null || doc >= ctx.docBase + ctx.reader().maxDoc());
            assert doc >= ctx.docBase;
            switch(numericType) {
                case LONG:
                case DATE:
                case INTEGER:
                    // Long, Date and Integer
                    longs = DocValues.getNumeric(ctx.reader(), fieldName);
                    break;
                case FLOAT:
                    // TODO: this bit flipping should probably be moved to tie-break in the PQ comparator
                    longs = new FilterNumericDocValues(DocValues.getNumeric(ctx.reader(), fieldName)) {

                        @Override
                        public long longValue() throws IOException {
                            long bits = super.longValue();
                            if (bits < 0)
                                bits ^= 0x7fffffffffffffffL;
                            return bits;
                        }
                    };
                    break;
                case DOUBLE:
                    // TODO: this bit flipping should probably be moved to tie-break in the PQ comparator
                    longs = new FilterNumericDocValues(DocValues.getNumeric(ctx.reader(), fieldName)) {

                        @Override
                        public long longValue() throws IOException {
                            long bits = super.longValue();
                            if (bits < 0)
                                bits ^= 0x7fffffffffffffffL;
                            return bits;
                        }
                    };
                    break;
                default:
                    throw new AssertionError("Unexpected type: " + numericType);
            }
        }
        int valuesDocID = longs.docID();
        if (valuesDocID < doc - ctx.docBase) {
            valuesDocID = longs.advance(doc - ctx.docBase);
        }
        if (valuesDocID == doc - ctx.docBase) {
            hashTable.add(doc, longs.longValue(), 1);
        } else {
            ++missingCount;
        }
    }
    // 2. select top-k facet values
    final int pqSize = limit < 0 ? hashTable.size : Math.min(offset + limit, hashTable.size);
    final PriorityQueue<Entry> pq;
    if (FacetParams.FACET_SORT_COUNT.equals(sort) || FacetParams.FACET_SORT_COUNT_LEGACY.equals(sort)) {
        pq = new PriorityQueue<Entry>(pqSize) {

            @Override
            protected boolean lessThan(Entry a, Entry b) {
                if (a.count < b.count || (a.count == b.count && a.bits > b.bits)) {
                    return true;
                } else {
                    return false;
                }
            }
        };
    } else {
        pq = new PriorityQueue<Entry>(pqSize) {

            @Override
            protected boolean lessThan(Entry a, Entry b) {
                return a.bits > b.bits;
            }
        };
    }
    Entry e = null;
    for (int i = 0; i < hashTable.bits.length; ++i) {
        if (hashTable.counts[i] >= mincount) {
            if (e == null) {
                e = new Entry();
            }
            e.bits = hashTable.bits[i];
            e.count = hashTable.counts[i];
            e.docID = hashTable.docIDs[i];
            e = pq.insertWithOverflow(e);
        }
    }
    // 4. build the NamedList
    final ValueSource vs = ft.getValueSource(sf, null);
    final NamedList<Integer> result = new NamedList<>();
    // to be merged with terms from the terms dict
    if (!zeros || FacetParams.FACET_SORT_COUNT.equals(sort) || FacetParams.FACET_SORT_COUNT_LEGACY.equals(sort)) {
        // Only keep items we're interested in
        final Deque<Entry> counts = new ArrayDeque<>();
        while (pq.size() > offset) {
            counts.addFirst(pq.pop());
        }
        // Entries from the PQ first, then using the terms dictionary
        for (Entry entry : counts) {
            final int readerIdx = ReaderUtil.subIndex(entry.docID, leaves);
            final FunctionValues values = vs.getValues(Collections.emptyMap(), leaves.get(readerIdx));
            result.add(values.strVal(entry.docID - leaves.get(readerIdx).docBase), entry.count);
        }
        if (zeros && (limit < 0 || result.size() < limit)) {
            // need to merge with the term dict
            if (!sf.indexed() && !sf.hasDocValues()) {
                throw new IllegalStateException("Cannot use " + FacetParams.FACET_MINCOUNT + "=0 on field " + sf.getName() + " which is neither indexed nor docValues");
            }
            // Add zeros until there are limit results
            final Set<String> alreadySeen = new HashSet<>();
            while (pq.size() > 0) {
                Entry entry = pq.pop();
                final int readerIdx = ReaderUtil.subIndex(entry.docID, leaves);
                final FunctionValues values = vs.getValues(Collections.emptyMap(), leaves.get(readerIdx));
                alreadySeen.add(values.strVal(entry.docID - leaves.get(readerIdx).docBase));
            }
            for (int i = 0; i < result.size(); ++i) {
                alreadySeen.add(result.getName(i));
            }
            final Terms terms = searcher.getSlowAtomicReader().terms(fieldName);
            if (terms != null) {
                final String prefixStr = TrieField.getMainValuePrefix(ft);
                final BytesRef prefix;
                if (prefixStr != null) {
                    prefix = new BytesRef(prefixStr);
                } else {
                    prefix = new BytesRef();
                }
                final TermsEnum termsEnum = terms.iterator();
                BytesRef term;
                switch(termsEnum.seekCeil(prefix)) {
                    case FOUND:
                    case NOT_FOUND:
                        term = termsEnum.term();
                        break;
                    case END:
                        term = null;
                        break;
                    default:
                        throw new AssertionError();
                }
                final CharsRefBuilder spare = new CharsRefBuilder();
                for (int skipped = hashTable.size; skipped < offset && term != null && StringHelper.startsWith(term, prefix); ) {
                    ft.indexedToReadable(term, spare);
                    final String termStr = spare.toString();
                    if (!alreadySeen.contains(termStr)) {
                        ++skipped;
                    }
                    term = termsEnum.next();
                }
                for (; term != null && StringHelper.startsWith(term, prefix) && (limit < 0 || result.size() < limit); term = termsEnum.next()) {
                    ft.indexedToReadable(term, spare);
                    final String termStr = spare.toString();
                    if (!alreadySeen.contains(termStr)) {
                        result.add(termStr, 0);
                    }
                }
            }
        }
    } else {
        // => Merge the PQ and the terms dictionary on the fly
        if (!sf.indexed()) {
            throw new IllegalStateException("Cannot use " + FacetParams.FACET_SORT + "=" + FacetParams.FACET_SORT_INDEX + " on a field which is not indexed");
        }
        final Map<String, Integer> counts = new HashMap<>();
        while (pq.size() > 0) {
            final Entry entry = pq.pop();
            final int readerIdx = ReaderUtil.subIndex(entry.docID, leaves);
            final FunctionValues values = vs.getValues(Collections.emptyMap(), leaves.get(readerIdx));
            counts.put(values.strVal(entry.docID - leaves.get(readerIdx).docBase), entry.count);
        }
        final Terms terms = searcher.getSlowAtomicReader().terms(fieldName);
        if (terms != null) {
            final String prefixStr = TrieField.getMainValuePrefix(ft);
            final BytesRef prefix;
            if (prefixStr != null) {
                prefix = new BytesRef(prefixStr);
            } else {
                prefix = new BytesRef();
            }
            final TermsEnum termsEnum = terms.iterator();
            BytesRef term;
            switch(termsEnum.seekCeil(prefix)) {
                case FOUND:
                case NOT_FOUND:
                    term = termsEnum.term();
                    break;
                case END:
                    term = null;
                    break;
                default:
                    throw new AssertionError();
            }
            final CharsRefBuilder spare = new CharsRefBuilder();
            for (int i = 0; i < offset && term != null && StringHelper.startsWith(term, prefix); ++i) {
                term = termsEnum.next();
            }
            for (; term != null && StringHelper.startsWith(term, prefix) && (limit < 0 || result.size() < limit); term = termsEnum.next()) {
                ft.indexedToReadable(term, spare);
                final String termStr = spare.toString();
                Integer count = counts.get(termStr);
                if (count == null) {
                    count = 0;
                }
                result.add(termStr, count);
            }
        }
    }
    if (missing) {
        result.add(null, missingCount);
    }
    return result;
}

Example 3 with FilterNumericDocValues

use of org.apache.lucene.index.FilterNumericDocValues in project lucene-solr by apache.

the class ToParentBlockJoinSortField method getDoubleComparator.

private FieldComparator<?> getDoubleComparator(int numHits) {
    return new FieldComparator.DoubleComparator(numHits, getField(), (Double) missingValue) {

        @Override
        protected NumericDocValues getNumericDocValues(LeafReaderContext context, String field) throws IOException {
            SortedNumericDocValues sortedNumeric = DocValues.getSortedNumeric(context.reader(), field);
            final BlockJoinSelector.Type type = order ? BlockJoinSelector.Type.MAX : BlockJoinSelector.Type.MIN;
            final BitSet parents = parentFilter.getBitSet(context);
            final BitSet children = childFilter.getBitSet(context);
            if (children == null) {
                return DocValues.emptyNumeric();
            }
            return new FilterNumericDocValues(BlockJoinSelector.wrap(sortedNumeric, type, parents, children)) {

                @Override
                public long longValue() throws IOException {
                    // undo the numericutils sortability
                    return NumericUtils.sortableDoubleBits(super.longValue());
                }
            };
        }
    };
}

Example 4 with FilterNumericDocValues

use of org.apache.lucene.index.FilterNumericDocValues in project lucene-solr by apache.

the class SortedNumericSelector method wrap.

/** 
   * Wraps a multi-valued SortedNumericDocValues as a single-valued view, using the specified selector 
   * and numericType.
   */
public static NumericDocValues wrap(SortedNumericDocValues sortedNumeric, Type selector, SortField.Type numericType) {
    if (numericType != SortField.Type.INT && numericType != SortField.Type.LONG && numericType != SortField.Type.FLOAT && numericType != SortField.Type.DOUBLE) {
        throw new IllegalArgumentException("numericType must be a numeric type");
    }
    final NumericDocValues view;
    NumericDocValues singleton = DocValues.unwrapSingleton(sortedNumeric);
    if (singleton != null) {
        // it's actually single-valued in practice, but indexed as multi-valued,
        // so just sort on the underlying single-valued dv directly.
        // regardless of selector type, this optimization is safe!
        view = singleton;
    } else {
        switch(selector) {
            case MIN:
                view = new MinValue(sortedNumeric);
                break;
            case MAX:
                view = new MaxValue(sortedNumeric);
                break;
            default:
                throw new AssertionError();
        }
    }
    // undo the numericutils sortability
    switch(numericType) {
        case FLOAT:
            return new FilterNumericDocValues(view) {

                @Override
                public long longValue() throws IOException {
                    return NumericUtils.sortableFloatBits((int) in.longValue());
                }
            };
        case DOUBLE:
            return new FilterNumericDocValues(view) {

                @Override
                public long longValue() throws IOException {
                    return NumericUtils.sortableDoubleBits(in.longValue());
                }
            };
        default:
            return view;
    }
}

Example 5 with FilterNumericDocValues

use of org.apache.lucene.index.FilterNumericDocValues in project lucene-solr by apache.

the class ToParentBlockJoinSortField method getFloatComparator.

private FieldComparator<?> getFloatComparator(int numHits) {
    return new FieldComparator.FloatComparator(numHits, getField(), (Float) missingValue) {

        @Override
        protected NumericDocValues getNumericDocValues(LeafReaderContext context, String field) throws IOException {
            SortedNumericDocValues sortedNumeric = DocValues.getSortedNumeric(context.reader(), field);
            final BlockJoinSelector.Type type = order ? BlockJoinSelector.Type.MAX : BlockJoinSelector.Type.MIN;
            final BitSet parents = parentFilter.getBitSet(context);
            final BitSet children = childFilter.getBitSet(context);
            if (children == null) {
                return DocValues.emptyNumeric();
            }
            return new FilterNumericDocValues(BlockJoinSelector.wrap(sortedNumeric, type, parents, children)) {

                @Override
                public long longValue() throws IOException {
                    // undo the numericutils sortability
                    return NumericUtils.sortableFloatBits((int) super.longValue());
                }
            };
        }
    };
}