Solr源码分析(10)之Lucene的索引文件(5)
作者:互联网
2021SC@SDUSC
1. .dvd和.dvm文件
.dvm是存放了DocValue域的元数据,比如DocValue偏移量。
.dvd则存放了DocValue的数据。
在Solr4.8.0中,dvd以及dvm用到的Lucene编码格式是Lucene45DocValuesFormat。跟之前的文件格式类似,它分别包含Lucene45DocValuesProducer
和Lucene45DocValuesConsumer来实现该文件的读和写。
@Override
public DocValuesConsumer fieldsConsumer(SegmentWriteState state) throws IOException {
return new Lucene45DocValuesConsumer(state, DATA_CODEC, DATA_EXTENSION, META_CODEC, META_EXTENSION);
}
@Override
public DocValuesProducer fieldsProducer(SegmentReadState state) throws IOException {
return new Lucene45DocValuesProducer(state, DATA_CODEC, DATA_EXTENSION, META_CODEC, META_EXTENSION);
}
Lucene 4.5 DocValues format通过下面的策略对四种类型进行编码:
-
NUMERIC
- Delta-compressed(增量压缩):表示文档值的整数写入一个16k大小的block。在每个block中,最小的值被编码,每个入口都是最小值的一个增量。所有这些增量都使用位压缩方法。更多信息见下文BlockPackedWriter。
- Table-compressed(表压缩):当不同数值的数量非常小(<256)时,或者在文档值序列中有没有使用gap值,solr会使用一个查找表替代。每一个文档值的入口都被替代为表上的序号,这些序号同样适用位压缩的方法压缩为PackedInts格式。
- GCD-compressed(最大公约数压缩):当所有数字分享同一个除数,最大共同分母(reatest common denominator,GCD) 会被计算出来,并且商会使用Delta-compressed 策略存储起来。
-
BINARY
- Fixed-width Binary:使用一个固定长度的,很大的拼接起来的位数组。每一个文档值都可以直接用docID * length 得到。
- Variable-width Binary:同样是一个很大的拼接起来的位数组,不过加入每篇文档的结束地址。这些地址从16k大小的块的起始位置被写入,并且每个入口具有平均长度的增量。对每篇文档,与增量之间的偏差(实际-平均)会记录下来。
- Prefix-compressed Binary:值会写入16(字节)大小的chunk中,其中第一个值被完整地记录下来,而其他值分享前缀。Chunk的地址被写入16k大小的block中。从block的起始位置开始写,对每个入口使用平均值作为增量。对每篇文档,与增量之间的偏差(实际-平均)会记录下来。
-
Sorted:
- 使用Prefix-compressed Binary压缩法实现一个从序号到重复的term的映射,同时所有文档的序号使用上面出现的Numeric压缩策略
-
SortedSet:
- 使用Prefix-compressed Binary压缩法实现一个从序号到重复的term的映射,同时一个序号的列表和所有文档在这个列表上的索引使用上面出现的Numeric压缩策略。
1.1 .dvm和.dvd文件格式
首先来介绍下.dvm的文件格式:
. dvm的文件结构分为好多层:
-
第一层:.dvm 由Header,NumFields,Footer
- Header和Footer跟之前相同
- NumFields 个的Entry,Entry为入口。
-
第二层:Entry具有四种类型,NumericEntry | BinaryEntry | SortedEntry | SortedSetEntry
-
第三层:
- NumericEntry:有三种类型GCDNumericEntry | TableNumericEntry | DeltaNumericEntry
- GCDNumericEntry: 包含NumericHeader,MinValue,GCD三部分
- TableNumericEntry:包含NumericHeader,TableSize,Int64TableSize三部分
- DeltaNumericEntry:包含NumericHeader
- BinaryEntry: 有三种类型FixedBinaryEntry | VariableBinaryEntry | PrefixBinaryEntry
- FixedBinaryEntry: 包含BinaryHeader
- VariableBinaryEntry:包含BinaryHeader,AddressOffset,PackedVersion,BlockSize四部分
- PrefixBinaryEntry: 包含BinaryHeader,AddressInterval,AddressOffset,PackedVersion,BlockSize五部分
- NumericEntry:有三种类型GCDNumericEntry | TableNumericEntry | DeltaNumericEntry
-
SortedEntry: 包含FieldNumber,EntryType,BinaryEntry,NumericEntry
-
SortedSetEntry: 包含EntryType,BinaryEntry,NumericEntry,NumericEntry
同样.dvd 文件具有好几层结构:
-
第一层:Header,<NumericData | BinaryData | SortedData>NumFields,Footer 与dvm类似,NumFields个Data(SortedData,BinaryData,NumericData其中一个)
-
第二层:
- NumericData:DeltaCompressedNumerics | TableCompressedNumerics | GCDCompressedNumerics对应上文讲到的Numeric的压缩方式
- BinaryData:ByteDataLength,Addresses
- SortedData:FST
-
第三层:
- DeltaCompressedNumerics:BlockPackedInts(blockSize=16k)
- TableCompressedNumerics:PackedInts
- GCDCompressedNumerics: BlockPackedInts(blockSize=16k)
- Addresses:MonotonicBlockPackedInts(blockSize=16k)
-
SortedSet入口储存了BinaryData中的序号的列表,使用一个增长的vLong类型的序列,并用差值编码。
1.2 .dvm和.dvd代码实现
前文讲到Lucene45DocValuesFormat分别包含Lucene45DocValuesProducer和Lucene45DocValuesConsumer来实现该文件的读和写,那么本节内容主要以Lucene45DocValuesProducer为例来学习下dvm和dvd。
首先学习下Lucene45DocValuesProducer的初始化:主要作用是读取.dvm文件和.dvd流。其中在读取.dvm文件过程中,Lucene45DocValuesProducer调用了readFields(in, state.fieldInfos)来获取入口信息。
protected Lucene45DocValuesProducer(SegmentReadState state, String dataCodec, String dataExtension, String metaCodec, String metaExtension) throws IOException {
//.dvm文件名
String metaName = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, metaExtension);
// read in the entries from the metadata file.
//打开.dvm并获取检验和,获取文件流,
ChecksumIndexInput in = state.directory.openChecksumInput(metaName, state.context);
//获取segment的document个数
this.maxDoc = state.segmentInfo.getDocCount();
boolean success = false;
try {
//获取.dvm header
version = CodecUtil.checkHeader(in, metaCodec,
Lucene45DocValuesFormat.VERSION_START,
Lucene45DocValuesFormat.VERSION_CURRENT);
numerics = new HashMap<>();
ords = new HashMap<>();
ordIndexes = new HashMap<>();
binaries = new HashMap<>();
sortedSets = new HashMap<>();
//读取NumFields个<Entry>
readFields(in, state.fieldInfos);
//加入Footer
if (version >= Lucene45DocValuesFormat.VERSION_CHECKSUM) {
CodecUtil.checkFooter(in);
} else {
CodecUtil.checkEOF(in);
}
success = true;
} finally {
if (success) {
IOUtils.close(in);
} else {
IOUtils.closeWhileHandlingException(in);
}
}
success = false;
try {
//.dvd文件名
String dataName = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, dataExtension);
//打开.dvd文件
data = state.directory.openInput(dataName, state.context);
//获取.dvd header
final int version2 = CodecUtil.checkHeader(data, dataCodec,
Lucene45DocValuesFormat.VERSION_START,
Lucene45DocValuesFormat.VERSION_CURRENT);
if (version != version2) {
throw new CorruptIndexException("Format versions mismatch");
}
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(this.data);
}
}
//估算类的大小,也就是估算.dvd流的大小
ramBytesUsed = new AtomicLong(RamUsageEstimator.shallowSizeOfInstance(getClass()));
}
readFields(in, state.fieldInfos)主要是读取EntryType,根据它的值来选择哪种方式来读取后续的Entry信息,
函数中涉及了以下几个方式:
- Numeric类型readNumericEntry()
2.BinaryEntry类型readBinaryEntry()
3.SortedSetEntry类型readSortedField()
4.SortedSetEntry类型readSortedSetEntry(),同时在该类型下,readFields还分别调用了readSortedSetFieldWithAddresses和readSortedField
private void readFields(IndexInput meta, FieldInfos infos) throws IOException {
//读取Entry的编号,如果编号为-1,表示这是最后一个Entry。
int fieldNumber = meta.readVInt();
while (fieldNumber != -1) {
// check should be: infos.fieldInfo(fieldNumber) != null, which incorporates negative check
// but docvalues updates are currently buggy here (loading extra stuff, etc): LUCENE-5616
if (fieldNumber < 0) {
// trickier to validate more: because we re-use for norms, because we use multiple entries
// for "composite" types like sortedset, etc.
throw new CorruptIndexException("Invalid field number: " + fieldNumber + " (resource=" + meta + ")");
}
//读取EntryType,以此来区分Entry的类型,0表示NUMERICENTRY,1表示BINARYENTRY,2表示SORTEDENTRY,3表示SORTED_SETENTRY
byte type = meta.readByte();
if (type == Lucene45DocValuesFormat.NUMERIC) {
//获取具体的NumericEntry内容,并放入以编号为键,NumericEntry为值的map中
numerics.put(fieldNumber, readNumericEntry(meta));
} else if (type == Lucene45DocValuesFormat.BINARY) {
//获取具体的BinaryEntry内容,并放入以编号为键,BinaryEntry为值的map中
BinaryEntry b = readBinaryEntry(meta);
binaries.put(fieldNumber, b);
} else if (type == Lucene45DocValuesFormat.SORTED) {
//读取SortedEntry
readSortedField(fieldNumber, meta, infos);
} else if (type == Lucene45DocValuesFormat.SORTED_SET) {
//读取SortedSetEntry,并放入以编号为键,SortedSetEntry为值的map中
SortedSetEntry ss = readSortedSetEntry(meta);
sortedSets.put(fieldNumber, ss);
//标准的存储有序的集合是否通过address的间接转换,SORTED_SET_WITH_ADDRESSES是docid->address>ord映射
if (ss.format == SORTED_SET_WITH_ADDRESSES) {
readSortedSetFieldWithAddresses(fieldNumber, meta, infos);
//SORTED_SET_SINGLE_VALUED_SORTED 只存储docid->ord的值
} else if (ss.format == SORTED_SET_SINGLE_VALUED_SORTED) {
if (meta.readVInt() != fieldNumber) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
if (meta.readByte() != Lucene45DocValuesFormat.SORTED) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
readSortedField(fieldNumber, meta, infos);
} else {
throw new AssertionError();
}
} else {
throw new CorruptIndexException("invalid type: " + type + ", resource=" + meta);
}
//读取下一个Entry
fieldNumber = meta.readVInt();
}
}
- readNumericEntry()
static NumericEntry readNumericEntry(IndexInput meta) throws IOException {
NumericEntry entry = new NumericEntry();
entry.format = meta.readVInt(); //NumericType,Numeric的编码方式
entry.missingOffset = meta.readLong(); //MissingOffset表示该field在哪个document中缺失,如果为-1表示没有document缺失字段
entry.packedIntsVersion = meta.readVInt(); //PackedVersion 打包整数的version
entry.offset = meta.readLong(); //DataOffset 指向.dvd文件中数据起始位置的指针
entry.count = meta.readVLong(); //Count 已写的值的个数
entry.blockSize = meta.readVInt(); //BlockSize 已打包的整数的大小
switch(entry.format) {
case GCD_COMPRESSED: //GCD-compressed(最大公约数压缩)
entry.minValue = meta.readLong(); //MinValue
entry.gcd = meta.readLong(); //GCD
break;
case TABLE_COMPRESSED: //Table-compressed(表压缩)
if (entry.count > Integer.MAX_VALUE) {
throw new CorruptIndexException("Cannot use TABLE_COMPRESSED with more than MAX_VALUE values, input=" + meta);
}
final int uniqueValues = meta.readVInt(); //TableSize
if (uniqueValues > 256) { //TableSize必须小于256
throw new CorruptIndexException("TABLE_COMPRESSED cannot have more than 256 distinct values, input=" + meta);
}
entry.table = new long[uniqueValues]; //TableSize个Long
for (int i = 0; i < uniqueValues; ++i) {
entry.table[i] = meta.readLong();
}
break;
case DELTA_COMPRESSED: //Delta-compressed(增量压缩)
break;
default:
throw new CorruptIndexException("Unknown format: " + entry.format + ", input=" + meta);
}
return entry;
}
- BinaryEntry()
static BinaryEntry readBinaryEntry(IndexInput meta) throws IOException {
BinaryEntry entry = new BinaryEntry();
entry.format = meta.readVInt(); //BinaryType类型
entry.missingOffset = meta.readLong(); //缺失表示,同NuericEntry
entry.minLength = meta.readVInt(); //存储Binary 类型的值的位数组的长度的最小值和最大值。
//如果这两个值是相等的,那么所有的值都是固定的大小,
//并且可以通过DataOffset + (docID * length)计算出来。
//否则,Binary的值是不定长的
entry.maxLength = meta.readVInt();
entry.count = meta.readVLong();
entry.offset = meta.readLong(); //实际二进制数的偏移
switch(entry.format) {
case BINARY_FIXED_UNCOMPRESSED: //Fixed-width Binary
break;
case BINARY_PREFIX_COMPRESSED: //Variable-width Binary
entry.addressInterval = meta.readVInt();
entry.addressesOffset = meta.readLong();
entry.packedIntsVersion = meta.readVInt();
entry.blockSize = meta.readVInt();
break;
case BINARY_VARIABLE_UNCOMPRESSED: //Prefix-compressed Binary
entry.addressesOffset = meta.readLong();
entry.packedIntsVersion = meta.readVInt();
entry.blockSize = meta.readVInt();
break;
default:
throw new CorruptIndexException("Unknown format: " + entry.format + ", input=" + meta);
}
return entry;
}
- readSortedSetFieldWithAddresses()
private void readSortedSetFieldWithAddresses(int fieldNumber, IndexInput meta, FieldInfos infos) throws IOException {
// sortedset = binary + numeric (addresses) + ordIndex
if (meta.readVInt() != fieldNumber) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
if (meta.readByte() != Lucene45DocValuesFormat.BINARY) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
BinaryEntry b = readBinaryEntry(meta);
binaries.put(fieldNumber, b);
if (meta.readVInt() != fieldNumber) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
if (meta.readByte() != Lucene45DocValuesFormat.NUMERIC) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
NumericEntry n1 = readNumericEntry(meta);
ords.put(fieldNumber, n1);
if (meta.readVInt() != fieldNumber) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
if (meta.readByte() != Lucene45DocValuesFormat.NUMERIC) {
throw new CorruptIndexException("sortedset entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
NumericEntry n2 = readNumericEntry(meta);
ordIndexes.put(fieldNumber, n2);
}
- readSortedField
private void readSortedField(int fieldNumber, IndexInput meta, FieldInfos infos) throws IOException {
// sorted = binary + numeric
if (meta.readVInt() != fieldNumber) {
throw new CorruptIndexException("sorted entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
if (meta.readByte() != Lucene45DocValuesFormat.BINARY) {
throw new CorruptIndexException("sorted entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
BinaryEntry b = readBinaryEntry(meta);
binaries.put(fieldNumber, b);
if (meta.readVInt() != fieldNumber) {
throw new CorruptIndexException("sorted entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
if (meta.readByte() != Lucene45DocValuesFormat.NUMERIC) {
throw new CorruptIndexException("sorted entry for field: " + fieldNumber + " is corrupt (resource=" + meta + ")");
}
NumericEntry n = readNumericEntry(meta);
ords.put(fieldNumber, n);
}
上文讲了.dvm文件的读取, 那么接下来学习下怎么对.dvd文件的读取。
标签:10,readVInt,new,Lucene,源码,meta,entry,fieldNumber,throw 来源: https://blog.csdn.net/hgdshbrjt/article/details/122147808