Package com.apple.foundationdb.record.lucene

Class LucenePrimaryKeySegmentIndexV1

java.lang.Object

com.apple.foundationdb.record.lucene.LucenePrimaryKeySegmentIndexV1

All Implemented Interfaces:

LucenePrimaryKeySegmentIndex

public class LucenePrimaryKeySegmentIndexV1
extends Object
implements LucenePrimaryKeySegmentIndex

Maintain a B-tree index of primary key to segment and doc id. This allows for efficient deleting of a document given that key, such as when doing index maintenance from an update. This works with any implementation of StoredFieldsWriter.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	LucenePrimaryKeySegmentIndexV1.StoredFieldsReaderSegmentInfo	Hook for getting back segment info during merge.

Nested classes/interfaces inherited from interface com.apple.foundationdb.record.lucene.LucenePrimaryKeySegmentIndex

LucenePrimaryKeySegmentIndex.DocumentIndexEntry

Constructor Summary

Constructors

Constructor	Description
LucenePrimaryKeySegmentIndexV1(FDBDirectory directory, Subspace subspace)

Method Summary

Modifier and Type	Method	Description
void	addOrDeletePrimaryKeyEntry(byte[] primaryKey, long segmentId, int docId, boolean add, String segmentName)	Add or delete the primary key/segment/docId from the index.
void	clearForSegment(String segmentName)	Clears all the primary key entries for a given segment name.
LucenePrimaryKeySegmentIndex.DocumentIndexEntry	findDocument(org.apache.lucene.index.DirectoryReader directoryReader, Tuple primaryKey)	Find document in index for direct delete.
List<String>	findSegments(Tuple primaryKey)	Return all the segments in which the given primary key appears.
CompletableFuture<Boolean>	hasPrimaryKeyAsync(Tuple primaryKey)	Check whether the index has an entry for the given primary key.
List<List<Object>>	readAllEntries()	Get all stored primary key index entries.
org.apache.lucene.codecs.StoredFieldsWriter	wrapFieldsWriter(org.apache.lucene.codecs.StoredFieldsWriter storedFieldsWriter, org.apache.lucene.index.SegmentInfo si)	Hook the fields writer to also record primary keys in index.

Methods inherited from class java.lang.Object

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

Constructor Details

LucenePrimaryKeySegmentIndexV1

public LucenePrimaryKeySegmentIndexV1(@Nonnull
 FDBDirectory directory,
 @Nonnull
 Subspace subspace)

Method Details

readAllEntries

public List<List<Object>> readAllEntries()

Get all stored primary key index entries. Really only useful for small-scale debugging.

Specified by:

readAllEntries in interface LucenePrimaryKeySegmentIndex

Returns:

a list of Tuple-decoded key entries

findSegments

public List<String> findSegments(@Nonnull
 Tuple primaryKey)
                          throws IOException

Return all the segments in which the given primary key appears. Mostly for debug logging.

Specified by:

findSegments in interface LucenePrimaryKeySegmentIndex

Parameters:

primaryKey - the document's record's primary key

Returns:

a list of segment names or segment ids when apparently not associated with a name

Throws:

IOException

findDocument

@Nullable
public LucenePrimaryKeySegmentIndex.DocumentIndexEntry findDocument(@Nonnull
 org.apache.lucene.index.DirectoryReader directoryReader,
 @Nonnull
 Tuple primaryKey)
                                                             throws IOException

Description copied from interface: LucenePrimaryKeySegmentIndex

Find document in index for direct delete.

Specified by:

findDocument in interface LucenePrimaryKeySegmentIndex

Parameters:

directoryReader - a NRT reader

primaryKey - the document's record's primary key

Returns:

an entry with the leaf reader and document id in that segment or null if not found

Throws:

IOException

See Also:

• IndexWriter.tryDeleteDocument(org.apache.lucene.index.IndexReader, int)

hasPrimaryKeyAsync

@Nonnull
public CompletableFuture<Boolean> hasPrimaryKeyAsync(@Nonnull
 Tuple primaryKey)

Description copied from interface: LucenePrimaryKeySegmentIndex

Check whether the index has an entry for the given primary key.

Specified by:

hasPrimaryKeyAsync in interface LucenePrimaryKeySegmentIndex

Parameters:

primaryKey - the document's record's primary key

Returns:

a future that completes with true if an entry was found for the primary key

wrapFieldsWriter

@Nonnull
public org.apache.lucene.codecs.StoredFieldsWriter wrapFieldsWriter(@Nonnull
 org.apache.lucene.codecs.StoredFieldsWriter storedFieldsWriter,
 @Nonnull
 org.apache.lucene.index.SegmentInfo si)
                                                             throws IOException

Hook the fields writer to also record primary keys in index.

Parameters:

storedFieldsWriter - normal field writer

si - segment info for current writer

Returns:

a wrapped writer

Throws:

IOException - thrown by called methods

addOrDeletePrimaryKeyEntry

public void addOrDeletePrimaryKeyEntry(@Nonnull
 byte[] primaryKey,
 long segmentId,
 int docId,
 boolean add,
 String segmentName)

Description copied from interface: LucenePrimaryKeySegmentIndex

Add or delete the primary key/segment/docId from the index.

Specified by:

addOrDeletePrimaryKeyEntry in interface LucenePrimaryKeySegmentIndex

Parameters:

primaryKey - the primary ey of the record

segmentId - the id of the segment (see FDBDirectory.primaryKeySegmentId(java.lang.String, boolean))

docId - the document id within the segment

add - whether to add (true) or delete (false) the entry

segmentName - name associated with the segment, for logging

clearForSegment

public void clearForSegment(String segmentName)

Description copied from interface: LucenePrimaryKeySegmentIndex

Clears all the primary key entries for a given segment name.

Specified by:

clearForSegment in interface LucenePrimaryKeySegmentIndex

Parameters:

segmentName - the name of the segment to clear out