Package org.apache.lucene.index.memory

Class MemoryIndex.Info

java.lang.Object

org.apache.lucene.index.memory.MemoryIndex.Info

Enclosing class:

MemoryIndex

private final class MemoryIndex.Info
extends java.lang.Object

Index data structure for a field; contains the tokenized term texts and their positions.

Field Summary

Fields

Modifier and Type	Field	Description
private MemoryIndex.BinaryDocValuesProducer	binaryProducer
private FieldInfo	fieldInfo
private int	lastOffset	the last offset encountered in this field for multi field support
private int	lastPosition	the last position encountered in this field for multi field support
private byte[]	maxPackedValue
private int	maxTermFrequency
private byte[]	minPackedValue
private java.lang.Long	norm
private MemoryIndex.NumericDocValuesProducer	numericProducer
private int	numOverlapTokens	Number of overlapping tokens for this field
private int	numTokens	Number of added tokens for this field
private BytesRef[]	pointValues
private int	pointValuesCount
private boolean	preparedDocValuesAndPointValues
private MemoryIndex.SliceByteStartArray	sliceArray
private int[]	sortedTerms	Terms sorted ascending by term text; computed on demand
private long	sumTotalTermFreq
private BytesRefHash	terms	Term strings and their positions for this field: Map <String termText, ArrayIntList positions>

Constructor Summary

Constructors

Modifier	Constructor	Description
private	Info(FieldInfo fieldInfo, ByteBlockPool byteBlockPool)

Method Summary

Modifier and Type	Method	Description
(package private) void	freeze()
(package private) NumericDocValues	getNormDocValues()
(package private) void	prepareDocValuesAndPointValues()
(package private) void	sortTerms()	Sorts hashed terms into ascending order, reusing memory along the way.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

fieldInfo

private FieldInfo fieldInfo

norm

private java.lang.Long norm

terms

private BytesRefHash terms

Term strings and their positions for this field: Map <String termText, ArrayIntList positions>

sliceArray

private MemoryIndex.SliceByteStartArray sliceArray

sortedTerms

private transient int[] sortedTerms

Terms sorted ascending by term text; computed on demand

numTokens

private int numTokens

Number of added tokens for this field

numOverlapTokens

private int numOverlapTokens

Number of overlapping tokens for this field

sumTotalTermFreq

private long sumTotalTermFreq

maxTermFrequency

private int maxTermFrequency

lastPosition

private int lastPosition

the last position encountered in this field for multi field support

lastOffset

private int lastOffset

the last offset encountered in this field for multi field support

binaryProducer

private MemoryIndex.BinaryDocValuesProducer binaryProducer

numericProducer

private MemoryIndex.NumericDocValuesProducer numericProducer

preparedDocValuesAndPointValues

private boolean preparedDocValuesAndPointValues

pointValues

private BytesRef[] pointValues

minPackedValue

private byte[] minPackedValue

maxPackedValue

private byte[] maxPackedValue

pointValuesCount

private int pointValuesCount

Constructor Detail

Info

private Info(FieldInfo fieldInfo,
             ByteBlockPool byteBlockPool)

Method Detail

freeze

void freeze()

sortTerms

void sortTerms()

Sorts hashed terms into ascending order, reusing memory along the way. Note that sorting is lazily delayed until required (often it's not required at all). If a sorted view is required then hashing + sort + binary search is still faster and smaller than TreeMap usage (which would be an alternative and somewhat more elegant approach, apart from more sophisticated Tries / prefix trees).

prepareDocValuesAndPointValues

void prepareDocValuesAndPointValues()

getNormDocValues

NumericDocValues getNormDocValues()