Esempio n. 1
0
        public virtual void Merge(MergeState mergeState, IndexOptions indexOptions, TermsEnum termsEnum)
        {
            BytesRef term;

            Debug.Assert(termsEnum != null);
            long        sumTotalTermFreq         = 0;
            long        sumDocFreq               = 0;
            long        sumDFsinceLastAbortCheck = 0;
            FixedBitSet visitedDocs              = new FixedBitSet(mergeState.SegmentInfo.DocCount);

            if (indexOptions == IndexOptions.DOCS_ONLY)
            {
                if (docsEnum == null)
                {
                    docsEnum = new MappingMultiDocsEnum();
                }
                docsEnum.MergeState = mergeState;

                MultiDocsEnum docsEnumIn = null;

                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    docsEnumIn = (MultiDocsEnum)termsEnum.Docs(null, docsEnumIn, DocsFlags.NONE);
                    if (docsEnumIn != null)
                    {
                        docsEnum.Reset(docsEnumIn);
                        PostingsConsumer postingsConsumer = StartTerm(term);
                        TermStats        stats            = postingsConsumer.Merge(mergeState, indexOptions, docsEnum, visitedDocs);
                        if (stats.DocFreq > 0)
                        {
                            FinishTerm(term, stats);
                            sumTotalTermFreq         += stats.DocFreq;
                            sumDFsinceLastAbortCheck += stats.DocFreq;
                            sumDocFreq += stats.DocFreq;
                            if (sumDFsinceLastAbortCheck > 60000)
                            {
                                mergeState.CheckAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                                sumDFsinceLastAbortCheck = 0;
                            }
                        }
                    }
                }
            }
            else if (indexOptions == IndexOptions.DOCS_AND_FREQS)
            {
                if (docsAndFreqsEnum == null)
                {
                    docsAndFreqsEnum = new MappingMultiDocsEnum();
                }
                docsAndFreqsEnum.MergeState = mergeState;

                MultiDocsEnum docsAndFreqsEnumIn = null;

                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    docsAndFreqsEnumIn = (MultiDocsEnum)termsEnum.Docs(null, docsAndFreqsEnumIn);
                    Debug.Assert(docsAndFreqsEnumIn != null);
                    docsAndFreqsEnum.Reset(docsAndFreqsEnumIn);
                    PostingsConsumer postingsConsumer = StartTerm(term);
                    TermStats        stats            = postingsConsumer.Merge(mergeState, indexOptions, docsAndFreqsEnum, visitedDocs);
                    if (stats.DocFreq > 0)
                    {
                        FinishTerm(term, stats);
                        sumTotalTermFreq         += stats.TotalTermFreq;
                        sumDFsinceLastAbortCheck += stats.DocFreq;
                        sumDocFreq += stats.DocFreq;
                        if (sumDFsinceLastAbortCheck > 60000)
                        {
                            mergeState.CheckAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                            sumDFsinceLastAbortCheck = 0;
                        }
                    }
                }
            }
            else if (indexOptions == IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
            {
                if (postingsEnum == null)
                {
                    postingsEnum = new MappingMultiDocsAndPositionsEnum();
                }
                postingsEnum.MergeState = mergeState;
                MultiDocsAndPositionsEnum postingsEnumIn = null;
                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    postingsEnumIn = (MultiDocsAndPositionsEnum)termsEnum.DocsAndPositions(null, postingsEnumIn, DocsAndPositionsFlags.PAYLOADS);
                    Debug.Assert(postingsEnumIn != null);
                    postingsEnum.Reset(postingsEnumIn);

                    PostingsConsumer postingsConsumer = StartTerm(term);
                    TermStats        stats            = postingsConsumer.Merge(mergeState, indexOptions, postingsEnum, visitedDocs);
                    if (stats.DocFreq > 0)
                    {
                        FinishTerm(term, stats);
                        sumTotalTermFreq         += stats.TotalTermFreq;
                        sumDFsinceLastAbortCheck += stats.DocFreq;
                        sumDocFreq += stats.DocFreq;
                        if (sumDFsinceLastAbortCheck > 60000)
                        {
                            mergeState.CheckAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                            sumDFsinceLastAbortCheck = 0;
                        }
                    }
                }
            }
            else
            {
                Debug.Assert(indexOptions == IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
                if (postingsEnum == null)
                {
                    postingsEnum = new MappingMultiDocsAndPositionsEnum();
                }
                postingsEnum.MergeState = mergeState;
                MultiDocsAndPositionsEnum postingsEnumIn = null;
                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    postingsEnumIn = (MultiDocsAndPositionsEnum)termsEnum.DocsAndPositions(null, postingsEnumIn);
                    Debug.Assert(postingsEnumIn != null);
                    postingsEnum.Reset(postingsEnumIn);

                    PostingsConsumer postingsConsumer = StartTerm(term);
                    TermStats        stats            = postingsConsumer.Merge(mergeState, indexOptions, postingsEnum, visitedDocs);
                    if (stats.DocFreq > 0)
                    {
                        FinishTerm(term, stats);
                        sumTotalTermFreq         += stats.TotalTermFreq;
                        sumDFsinceLastAbortCheck += stats.DocFreq;
                        sumDocFreq += stats.DocFreq;
                        if (sumDFsinceLastAbortCheck > 60000)
                        {
                            mergeState.CheckAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                            sumDFsinceLastAbortCheck = 0;
                        }
                    }
                }
            }
            Finish(indexOptions == IndexOptions.DOCS_ONLY ? -1 : sumTotalTermFreq, sumDocFreq, visitedDocs.Cardinality());
        }
Esempio n. 2
0
        /// <summary>
        /// Default merge impl </summary>
        public virtual void Merge(MergeState mergeState, FieldInfo.IndexOptions? indexOptions, TermsEnum termsEnum)
        {
            BytesRef term;
            Debug.Assert(termsEnum != null);
            long sumTotalTermFreq = 0;
            long sumDocFreq = 0;
            long sumDFsinceLastAbortCheck = 0;
            FixedBitSet visitedDocs = new FixedBitSet(mergeState.SegmentInfo.DocCount);

            if (indexOptions == FieldInfo.IndexOptions.DOCS_ONLY)
            {
                if (DocsEnum == null)
                {
                    DocsEnum = new MappingMultiDocsEnum();
                }
                DocsEnum.MergeState = mergeState;

                MultiDocsEnum docsEnumIn = null;

                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    docsEnumIn = (MultiDocsEnum)termsEnum.Docs(null, docsEnumIn, Index.DocsEnum.FLAG_NONE);
                    if (docsEnumIn != null)
                    {
                        DocsEnum.Reset(docsEnumIn);
                        PostingsConsumer postingsConsumer = StartTerm(term);
                        TermStats stats = postingsConsumer.Merge(mergeState, indexOptions, DocsEnum, visitedDocs);
                        if (stats.DocFreq > 0)
                        {
                            FinishTerm(term, stats);
                            sumTotalTermFreq += stats.DocFreq;
                            sumDFsinceLastAbortCheck += stats.DocFreq;
                            sumDocFreq += stats.DocFreq;
                            if (sumDFsinceLastAbortCheck > 60000)
                            {
                                mergeState.checkAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                                sumDFsinceLastAbortCheck = 0;
                            }
                        }
                    }
                }
            }
            else if (indexOptions == FieldInfo.IndexOptions.DOCS_AND_FREQS)
            {
                if (DocsAndFreqsEnum == null)
                {
                    DocsAndFreqsEnum = new MappingMultiDocsEnum();
                }
                DocsAndFreqsEnum.MergeState = mergeState;

                MultiDocsEnum docsAndFreqsEnumIn = null;

                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    docsAndFreqsEnumIn = (MultiDocsEnum)termsEnum.Docs(null, docsAndFreqsEnumIn);
                    Debug.Assert(docsAndFreqsEnumIn != null);
                    DocsAndFreqsEnum.Reset(docsAndFreqsEnumIn);
                    PostingsConsumer postingsConsumer = StartTerm(term);
                    TermStats stats = postingsConsumer.Merge(mergeState, indexOptions, DocsAndFreqsEnum, visitedDocs);
                    if (stats.DocFreq > 0)
                    {
                        FinishTerm(term, stats);
                        sumTotalTermFreq += stats.TotalTermFreq;
                        sumDFsinceLastAbortCheck += stats.DocFreq;
                        sumDocFreq += stats.DocFreq;
                        if (sumDFsinceLastAbortCheck > 60000)
                        {
                            mergeState.checkAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                            sumDFsinceLastAbortCheck = 0;
                        }
                    }
                }
            }
            else if (indexOptions == FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
            {
                if (PostingsEnum == null)
                {
                    PostingsEnum = new MappingMultiDocsAndPositionsEnum();
                }
                PostingsEnum.MergeState = mergeState;
                MultiDocsAndPositionsEnum postingsEnumIn = null;
                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    postingsEnumIn = (MultiDocsAndPositionsEnum)termsEnum.DocsAndPositions(null, postingsEnumIn, DocsAndPositionsEnum.FLAG_PAYLOADS);
                    Debug.Assert(postingsEnumIn != null);
                    PostingsEnum.Reset(postingsEnumIn);

                    PostingsConsumer postingsConsumer = StartTerm(term);
                    TermStats stats = postingsConsumer.Merge(mergeState, indexOptions, PostingsEnum, visitedDocs);
                    if (stats.DocFreq > 0)
                    {
                        FinishTerm(term, stats);
                        sumTotalTermFreq += stats.TotalTermFreq;
                        sumDFsinceLastAbortCheck += stats.DocFreq;
                        sumDocFreq += stats.DocFreq;
                        if (sumDFsinceLastAbortCheck > 60000)
                        {
                            mergeState.checkAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                            sumDFsinceLastAbortCheck = 0;
                        }
                    }
                }
            }
            else
            {
                Debug.Assert(indexOptions == FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
                if (PostingsEnum == null)
                {
                    PostingsEnum = new MappingMultiDocsAndPositionsEnum();
                }
                PostingsEnum.MergeState = mergeState;
                MultiDocsAndPositionsEnum postingsEnumIn = null;
                while ((term = termsEnum.Next()) != null)
                {
                    // We can pass null for liveDocs, because the
                    // mapping enum will skip the non-live docs:
                    postingsEnumIn = (MultiDocsAndPositionsEnum)termsEnum.DocsAndPositions(null, postingsEnumIn);
                    Debug.Assert(postingsEnumIn != null);
                    PostingsEnum.Reset(postingsEnumIn);

                    PostingsConsumer postingsConsumer = StartTerm(term);
                    TermStats stats = postingsConsumer.Merge(mergeState, indexOptions, PostingsEnum, visitedDocs);
                    if (stats.DocFreq > 0)
                    {
                        FinishTerm(term, stats);
                        sumTotalTermFreq += stats.TotalTermFreq;
                        sumDFsinceLastAbortCheck += stats.DocFreq;
                        sumDocFreq += stats.DocFreq;
                        if (sumDFsinceLastAbortCheck > 60000)
                        {
                            mergeState.checkAbort.Work(sumDFsinceLastAbortCheck / 5.0);
                            sumDFsinceLastAbortCheck = 0;
                        }
                    }
                }
            }
            Finish(indexOptions == FieldInfo.IndexOptions.DOCS_ONLY ? -1 : sumTotalTermFreq, sumDocFreq, visitedDocs.Cardinality());
        }