org.apache.lucene.codecs.lucene40.values

Class Bytes

java.lang.Object

org.apache.lucene.codecs.lucene40.values.Bytes

public final class Bytes
extends Object

Provides concrete Writer/Reader implementations for byte[] value per document. There are 6 package-private default implementations of this, for all combinations of Bytes.Mode.DEREF/Bytes.Mode.STRAIGHT x fixed-length/variable-length.

NOTE: Currently the total amount of byte[] data stored (across a single segment) cannot exceed 2GB.

NOTE: Each byte[] must be <= 32768 bytes in length

WARNING: This API is experimental and might change in incompatible ways in the next release.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Bytes.Mode Defines the Writers store mode.

Method Summary

Methods

Modifier and Type	Method and Description
static DocValues	getValues(Directory dir, String id, Bytes.Mode mode, boolean fixedSize, int maxDoc, Comparator<BytesRef> sortComparator, IOContext context) Creates a new DocValues instance that provides either memory resident or iterative access to a per-document stored byte[] value.
static DocValuesConsumer	getWriter(Directory dir, String id, Bytes.Mode mode, boolean fixedSize, Comparator<BytesRef> sortComparator, Counter bytesUsed, IOContext context, float acceptableOverheadRatio) Creates a new byte[] Writer instances for the given directory.

Methods inherited from class java.lang.Object

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

Method Detail

getWriter

public static DocValuesConsumer getWriter(Directory dir,
                          String id,
                          Bytes.Mode mode,
                          boolean fixedSize,
                          Comparator<BytesRef> sortComparator,
                          Counter bytesUsed,
                          IOContext context,
                          float acceptableOverheadRatio)

Creates a new byte[] Writer instances for the given directory.

Parameters:

dir - the directory to write the values to

id - the id used to create a unique file name. Usually composed out of the segment name and a unique id per segment.

mode - the writers store mode

fixedSize - true if all bytes subsequently passed to the Writer will have the same length

sortComparator - BytesRef comparator used by sorted variants. If null BytesRef.getUTF8SortedAsUnicodeComparator() is used instead

bytesUsed - an AtomicLong instance to track the used bytes within the Writer. A call to DocValuesConsumer.finish(int) will release all internally used resources and frees the memory tracking reference.

acceptableOverheadRatio - how to trade space for speed. This option is only applicable for docvalues of type DocValues.Type.BYTES_FIXED_SORTED and DocValues.Type.BYTES_VAR_SORTED.

context - I/O Context

Returns:

a new Writer instance

See Also:

PackedInts.getReader(org.apache.lucene.store.DataInput)

getValues

public static DocValues getValues(Directory dir,
                  String id,
                  Bytes.Mode mode,
                  boolean fixedSize,
                  int maxDoc,
                  Comparator<BytesRef> sortComparator,
                  IOContext context)
                           throws IOException

Creates a new DocValues instance that provides either memory resident or iterative access to a per-document stored byte[] value. The returned DocValues instance will be initialized without consuming a significant amount of memory.

Parameters:

dir - the directory to load the DocValues from.

id - the file ID in the Directory to load the values from.

mode - the mode used to store the values

fixedSize - true iff the values are stored with fixed-size, otherwise false

maxDoc - the number of document values stored for the given ID

sortComparator - BytesRef comparator used by sorted variants. If null BytesRef.getUTF8SortedAsUnicodeComparator() is used instead

Returns:

an initialized DocValues instance.

Throws:

IOException - if an IOException occurs