mirror of https://github.com/apache/lucene.git
remove confusing dead code
git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/trunk@1147671 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Michael McCandless
parent
521af09c38
commit
50be8c0489
|
|
@ -1,110 +0,0 @@
|
|||
package org.apache.lucene.index;
|
||||
|
||||
/**
|
||||
* 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.
|
||||
*/
|
||||
|
||||
/** Remaps docIDs after a merge has completed, where the
|
||||
* merged segments had at least one deletion. This is used
|
||||
* to renumber the buffered deletes in IndexWriter when a
|
||||
* merge of segments with deletions commits. */
|
||||
|
||||
final class MergeDocIDRemapper {
|
||||
int[] starts; // used for binary search of mapped docID
|
||||
int[] newStarts; // starts, minus the deletes
|
||||
int[][] docMaps; // maps docIDs in the merged set
|
||||
int minDocID; // minimum docID that needs renumbering
|
||||
int maxDocID; // 1+ the max docID that needs renumbering
|
||||
int docShift; // total # deleted docs that were compacted by this merge
|
||||
|
||||
public MergeDocIDRemapper(SegmentInfos infos, int[][] docMaps, int[] delCounts, MergePolicy.OneMerge merge, int mergedDocCount) {
|
||||
this.docMaps = docMaps;
|
||||
SegmentInfo firstSegment = merge.segments.get(0);
|
||||
int i = 0;
|
||||
while(true) {
|
||||
SegmentInfo info = infos.info(i);
|
||||
if (info.equals(firstSegment))
|
||||
break;
|
||||
minDocID += info.docCount;
|
||||
i++;
|
||||
}
|
||||
|
||||
int numDocs = 0;
|
||||
for(int j=0;j<docMaps.length;i++,j++) {
|
||||
numDocs += infos.info(i).docCount;
|
||||
assert infos.info(i).equals(merge.segments.get(j));
|
||||
}
|
||||
maxDocID = minDocID + numDocs;
|
||||
|
||||
starts = new int[docMaps.length];
|
||||
newStarts = new int[docMaps.length];
|
||||
|
||||
starts[0] = minDocID;
|
||||
newStarts[0] = minDocID;
|
||||
for(i=1;i<docMaps.length;i++) {
|
||||
final int lastDocCount = merge.segments.get(i-1).docCount;
|
||||
starts[i] = starts[i-1] + lastDocCount;
|
||||
newStarts[i] = newStarts[i-1] + lastDocCount - delCounts[i-1];
|
||||
}
|
||||
docShift = numDocs - mergedDocCount;
|
||||
|
||||
// There are rare cases when docShift is 0. It happens
|
||||
// if you try to delete a docID that's out of bounds,
|
||||
// because the SegmentReader still allocates deletedDocs
|
||||
// and pretends it has deletions ... so we can't make
|
||||
// this assert here
|
||||
// assert docShift > 0;
|
||||
|
||||
// Make sure it all adds up:
|
||||
assert docShift == maxDocID - (newStarts[docMaps.length-1] + merge.segments.get(docMaps.length-1).docCount - delCounts[docMaps.length-1]);
|
||||
}
|
||||
|
||||
public int remap(int oldDocID) {
|
||||
if (oldDocID < minDocID)
|
||||
// Unaffected by merge
|
||||
return oldDocID;
|
||||
else if (oldDocID >= maxDocID)
|
||||
// This doc was "after" the merge, so simple shift
|
||||
return oldDocID - docShift;
|
||||
else {
|
||||
// Binary search to locate this document & find its new docID
|
||||
int lo = 0; // search starts array
|
||||
int hi = docMaps.length - 1; // for first element less
|
||||
|
||||
while (hi >= lo) {
|
||||
int mid = (lo + hi) >>> 1;
|
||||
int midValue = starts[mid];
|
||||
if (oldDocID < midValue)
|
||||
hi = mid - 1;
|
||||
else if (oldDocID > midValue)
|
||||
lo = mid + 1;
|
||||
else { // found a match
|
||||
while (mid+1 < docMaps.length && starts[mid+1] == midValue) {
|
||||
mid++; // scan to last match
|
||||
}
|
||||
if (docMaps[mid] != null)
|
||||
return newStarts[mid] + docMaps[mid][oldDocID-starts[mid]];
|
||||
else
|
||||
return newStarts[mid] + oldDocID-starts[mid];
|
||||
}
|
||||
}
|
||||
if (docMaps[hi] != null)
|
||||
return newStarts[hi] + docMaps[hi][oldDocID-starts[hi]];
|
||||
else
|
||||
return newStarts[hi] + oldDocID-starts[hi];
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -39,7 +39,6 @@ import org.apache.lucene.store.IndexInput;
|
|||
import org.apache.lucene.store.IndexOutput;
|
||||
import org.apache.lucene.util.Bits;
|
||||
import org.apache.lucene.util.IOUtils;
|
||||
import org.apache.lucene.util.MultiBits;
|
||||
import org.apache.lucene.util.ReaderUtil;
|
||||
|
||||
/**
|
||||
|
|
@ -513,7 +512,6 @@ final class SegmentMerger {
|
|||
mergeState.mergedDocCount = mergedDocs;
|
||||
|
||||
// Remap docIDs
|
||||
mergeState.delCounts = new int[mergeState.readerCount];
|
||||
mergeState.docMaps = new int[mergeState.readerCount][];
|
||||
mergeState.docBase = new int[mergeState.readerCount];
|
||||
mergeState.hasPayloadProcessorProvider = payloadProcessorProvider != null;
|
||||
|
|
@ -528,11 +526,10 @@ final class SegmentMerger {
|
|||
|
||||
final IndexReader reader = readers.get(i);
|
||||
|
||||
mergeState.delCounts[i] = reader.numDeletedDocs();
|
||||
mergeState.docBase[i] = docBase;
|
||||
docBase += reader.numDocs();
|
||||
inputDocBase += reader.maxDoc();
|
||||
if (mergeState.delCounts[i] != 0) {
|
||||
if (reader.hasDeletions()) {
|
||||
int delCount = 0;
|
||||
final Bits liveDocs = reader.getLiveDocs();
|
||||
assert liveDocs != null;
|
||||
|
|
@ -547,7 +544,7 @@ final class SegmentMerger {
|
|||
docMap[j] = newDocID++;
|
||||
}
|
||||
}
|
||||
assert delCount == mergeState.delCounts[i]: "reader delCount=" + mergeState.delCounts[i] + " vs recomputed delCount=" + delCount;
|
||||
assert delCount == reader.numDeletedDocs(): "reader delCount=" + reader.numDeletedDocs() + " vs recomputed delCount=" + delCount;
|
||||
}
|
||||
|
||||
if (payloadProcessorProvider != null) {
|
||||
|
|
@ -557,12 +554,6 @@ final class SegmentMerger {
|
|||
codec = segmentWriteState.segmentCodecs.codec();
|
||||
final FieldsConsumer consumer = codec.fieldsConsumer(segmentWriteState);
|
||||
try {
|
||||
// NOTE: this is silly, yet, necessary -- we create a
|
||||
// MultiBits as our skip docs only to have it broken
|
||||
// apart when we step through the docs enums in
|
||||
// MultiDocsEnum.
|
||||
mergeState.multiLiveDocs = new MultiBits(bits, bitsStarts, true);
|
||||
|
||||
consumer.merge(mergeState,
|
||||
new MultiFields(fields.toArray(Fields.EMPTY_ARRAY),
|
||||
slices.toArray(ReaderUtil.Slice.EMPTY_ARRAY)));
|
||||
|
|
@ -591,8 +582,6 @@ final class SegmentMerger {
|
|||
}
|
||||
perDocBitsStarts.add(docBase);
|
||||
if (!perDocSlices.isEmpty()) {
|
||||
mergeState.multiLiveDocs = new MultiBits(perDocBits, perDocBitsStarts,
|
||||
true);
|
||||
final PerDocConsumer docsConsumer = codec
|
||||
.docsConsumer(new PerDocWriteState(segmentWriteState));
|
||||
boolean success = false;
|
||||
|
|
@ -612,14 +601,6 @@ final class SegmentMerger {
|
|||
|
||||
private MergeState mergeState;
|
||||
|
||||
int[][] getDocMaps() {
|
||||
return mergeState.docMaps;
|
||||
}
|
||||
|
||||
int[] getDelCounts() {
|
||||
return mergeState.delCounts;
|
||||
}
|
||||
|
||||
public boolean getAnyNonBulkMerges() {
|
||||
assert matchedCount <= readers.size();
|
||||
return matchedCount != readers.size();
|
||||
|
|
|
|||
|
|
@ -26,7 +26,6 @@ import org.apache.lucene.index.MergePolicy;
|
|||
import org.apache.lucene.index.PayloadProcessorProvider.DirPayloadProcessor;
|
||||
import org.apache.lucene.index.PayloadProcessorProvider.PayloadProcessor;
|
||||
import org.apache.lucene.store.Directory;
|
||||
import org.apache.lucene.util.Bits;
|
||||
|
||||
/** Holds common state used during segment merging
|
||||
*
|
||||
|
|
@ -36,10 +35,8 @@ public class MergeState {
|
|||
public List<IndexReader> readers; // Readers being merged
|
||||
public int readerCount; // Number of readers being merged
|
||||
public int[][] docMaps; // Maps docIDs around deletions
|
||||
public int[] delCounts; // Deletion count per reader
|
||||
public int[] docBase; // New docID base per reader
|
||||
public int mergedDocCount; // Total # merged docs
|
||||
public Bits multiLiveDocs;
|
||||
public CheckAbort checkAbort;
|
||||
|
||||
// Updated per field;
|
||||
|
|
|
|||
|
|
@ -69,7 +69,9 @@ public abstract class TermsConsumer {
|
|||
MultiDocsEnum docsEnumIn = null;
|
||||
|
||||
while((term = termsEnum.next()) != null) {
|
||||
docsEnumIn = (MultiDocsEnum) termsEnum.docs(mergeState.multiLiveDocs, docsEnumIn);
|
||||
// We can pass null for liveDocs, because the
|
||||
// mapping enum will skip the non-live docs:
|
||||
docsEnumIn = (MultiDocsEnum) termsEnum.docs(null, docsEnumIn);
|
||||
if (docsEnumIn != null) {
|
||||
docsEnum.reset(docsEnumIn);
|
||||
final PostingsConsumer postingsConsumer = startTerm(term);
|
||||
|
|
@ -93,7 +95,9 @@ public abstract class TermsConsumer {
|
|||
postingsEnum.setMergeState(mergeState);
|
||||
MultiDocsAndPositionsEnum postingsEnumIn = null;
|
||||
while((term = termsEnum.next()) != null) {
|
||||
postingsEnumIn = (MultiDocsAndPositionsEnum) termsEnum.docsAndPositions(mergeState.multiLiveDocs, postingsEnumIn);
|
||||
// We can pass null for liveDocs, because the
|
||||
// mapping enum will skip the non-live docs:
|
||||
postingsEnumIn = (MultiDocsAndPositionsEnum) termsEnum.docsAndPositions(null, postingsEnumIn);
|
||||
if (postingsEnumIn != null) {
|
||||
postingsEnum.reset(postingsEnumIn);
|
||||
// set PayloadProcessor
|
||||
|
|
|
|||
Loading…
Reference in New Issue