InsaneReader(LeafReader in, String insaneField) {
super(in);
this.insaneField = insaneField;
ArrayList<FieldInfo> filteredInfos = new ArrayList<>();
for (FieldInfo fi : in.getFieldInfos()) {
if (fi.name.equals(insaneField)) {
filteredInfos.add(
new FieldInfo(
fi.name,
fi.number,
fi.hasVectors(),
fi.omitsNorms(),
fi.hasPayloads(),
fi.getIndexOptions(),
DocValuesType.NONE,
-1,
Collections.emptyMap(),
fi.getPointDimensionCount(),
fi.getPointNumBytes()));
} else {
filteredInfos.add(fi);
}
}
fieldInfos = new FieldInfos(filteredInfos.toArray(new FieldInfo[filteredInfos.size()]));
}
/** Constructs a new FieldInfos from an array of FieldInfo objects */
public FieldInfos(FieldInfo[] infos) {
boolean hasVectors = false;
boolean hasProx = false;
boolean hasPayloads = false;
boolean hasOffsets = false;
boolean hasFreq = false;
boolean hasNorms = false;
boolean hasDocValues = false;
for (FieldInfo info : infos) {
if (info.number < 0) {
throw new IllegalArgumentException(
"illegal field number: " + info.number + " for field " + info.name);
}
FieldInfo previous = byNumber.put(info.number, info);
if (previous != null) {
throw new IllegalArgumentException(
"duplicate field numbers: "
+ previous.name
+ " and "
+ info.name
+ " have: "
+ info.number);
}
previous = byName.put(info.name, info);
if (previous != null) {
throw new IllegalArgumentException(
"duplicate field names: "
+ previous.number
+ " and "
+ info.number
+ " have: "
+ info.name);
}
hasVectors |= info.hasVectors();
hasProx |= info.getIndexOptions().compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0;
hasFreq |= info.getIndexOptions() != IndexOptions.DOCS;
hasOffsets |=
info.getIndexOptions().compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS)
>= 0;
hasNorms |= info.hasNorms();
hasDocValues |= info.getDocValuesType() != DocValuesType.NONE;
hasPayloads |= info.hasPayloads();
}
this.hasVectors = hasVectors;
this.hasProx = hasProx;
this.hasPayloads = hasPayloads;
this.hasOffsets = hasOffsets;
this.hasFreq = hasFreq;
this.hasNorms = hasNorms;
this.hasDocValues = hasDocValues;
this.values = Collections.unmodifiableCollection(byNumber.values());
}
public FieldInfo add(FieldInfo fi) {
// IMPORTANT - reuse the field number if possible for consistent field numbers across segments
return addOrUpdateInternal(
fi.name,
fi.number,
fi.hasVectors(),
fi.omitsNorms(),
fi.hasPayloads(),
fi.getIndexOptions(),
fi.getDocValuesType());
}
@Override
public void write(
Directory directory,
SegmentInfo segmentInfo,
String segmentSuffix,
FieldInfos infos,
IOContext context)
throws IOException {
final String fileName =
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, Lucene46FieldInfosFormat.EXTENSION);
try (IndexOutput output = directory.createOutput(fileName, context)) {
CodecUtil.writeHeader(
output, Lucene46FieldInfosFormat.CODEC_NAME, Lucene46FieldInfosFormat.FORMAT_CURRENT);
output.writeVInt(infos.size());
for (FieldInfo fi : infos) {
IndexOptions indexOptions = fi.getIndexOptions();
byte bits = 0x0;
if (fi.hasVectors()) bits |= Lucene46FieldInfosFormat.STORE_TERMVECTOR;
if (fi.omitsNorms()) bits |= Lucene46FieldInfosFormat.OMIT_NORMS;
if (fi.hasPayloads()) bits |= Lucene46FieldInfosFormat.STORE_PAYLOADS;
if (fi.getIndexOptions() != IndexOptions.NONE) {
bits |= Lucene46FieldInfosFormat.IS_INDEXED;
assert indexOptions.compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0
|| !fi.hasPayloads();
if (indexOptions == IndexOptions.DOCS) {
bits |= Lucene46FieldInfosFormat.OMIT_TERM_FREQ_AND_POSITIONS;
} else if (indexOptions == IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) {
bits |= Lucene46FieldInfosFormat.STORE_OFFSETS_IN_POSTINGS;
} else if (indexOptions == IndexOptions.DOCS_AND_FREQS) {
bits |= Lucene46FieldInfosFormat.OMIT_POSITIONS;
}
}
output.writeString(fi.name);
output.writeVInt(fi.number);
output.writeByte(bits);
// pack the DV types in one byte
final byte dv = docValuesByte(fi.getDocValuesType());
final byte nrm = docValuesByte(fi.hasNorms() ? DocValuesType.NUMERIC : DocValuesType.NONE);
assert (dv & (~0xF)) == 0 && (nrm & (~0x0F)) == 0;
byte val = (byte) (0xff & ((nrm << 4) | dv));
output.writeByte(val);
output.writeLong(fi.getDocValuesGen());
output.writeStringStringMap(fi.attributes());
}
CodecUtil.writeFooter(output);
}
}
public void testMerge() throws IOException {
final Codec codec = Codec.getDefault();
final SegmentInfo si =
new SegmentInfo(
mergedDir,
Version.LATEST,
mergedSegment,
-1,
false,
codec,
Collections.emptyMap(),
StringHelper.randomId(),
new HashMap<>());
SegmentMerger merger =
new SegmentMerger(
Arrays.<CodecReader>asList(reader1, reader2),
si,
InfoStream.getDefault(),
mergedDir,
new FieldInfos.FieldNumbers(),
newIOContext(random(), new IOContext(new MergeInfo(-1, -1, false, -1))));
MergeState mergeState = merger.merge();
int docsMerged = mergeState.segmentInfo.maxDoc();
assertTrue(docsMerged == 2);
// Should be able to open a new SegmentReader against the new directory
SegmentReader mergedReader =
new SegmentReader(
new SegmentCommitInfo(mergeState.segmentInfo, 0, -1L, -1L, -1L),
newIOContext(random()));
assertTrue(mergedReader != null);
assertTrue(mergedReader.numDocs() == 2);
Document newDoc1 = mergedReader.document(0);
assertTrue(newDoc1 != null);
// There are 2 unstored fields on the document
assertTrue(
DocHelper.numFields(newDoc1) == DocHelper.numFields(doc1) - DocHelper.unstored.size());
Document newDoc2 = mergedReader.document(1);
assertTrue(newDoc2 != null);
assertTrue(
DocHelper.numFields(newDoc2) == DocHelper.numFields(doc2) - DocHelper.unstored.size());
PostingsEnum termDocs =
TestUtil.docs(
random(), mergedReader, DocHelper.TEXT_FIELD_2_KEY, new BytesRef("field"), null, 0);
assertTrue(termDocs != null);
assertTrue(termDocs.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int tvCount = 0;
for (FieldInfo fieldInfo : mergedReader.getFieldInfos()) {
if (fieldInfo.hasVectors()) {
tvCount++;
}
}
// System.out.println("stored size: " + stored.size());
assertEquals("We do not have 3 fields that were indexed with term vector", 3, tvCount);
Terms vector = mergedReader.getTermVectors(0).terms(DocHelper.TEXT_FIELD_2_KEY);
assertNotNull(vector);
assertEquals(3, vector.size());
TermsEnum termsEnum = vector.iterator();
int i = 0;
while (termsEnum.next() != null) {
String term = termsEnum.term().utf8ToString();
int freq = (int) termsEnum.totalTermFreq();
// System.out.println("Term: " + term + " Freq: " + freq);
assertTrue(DocHelper.FIELD_2_TEXT.indexOf(term) != -1);
assertTrue(DocHelper.FIELD_2_FREQS[i] == freq);
i++;
}
TestSegmentReader.checkNorms(mergedReader);
mergedReader.close();
}