org.apache.cassandra.db

Class ColumnIndex.Builder

java.lang.Object

org.apache.cassandra.db.ColumnIndex.Builder

Enclosing class:

ColumnIndex

public static class ColumnIndex.Builder
extends java.lang.Object

Help to create an index for a column family based on size of columns, and write said columns to disk.

Constructor Summary

Constructors

Constructor and Description
Builder(ColumnFamily cf, java.nio.ByteBuffer key, DataOutputPlus output)
Builder(ColumnFamily cf, java.nio.ByteBuffer key, DataOutputPlus output, OnDiskAtom.SerializerForWriting serializer)

Method Summary

Modifier and Type	Method and Description
void	add(OnDiskAtom column)
ColumnIndex	build()
ColumnIndex	build(ColumnFamily cf) Serializes the index into in-memory structure with all required components such as Bloom Filter, index block size, IndexInfo list
ColumnIndex	buildForCompaction(java.util.Iterator<OnDiskAtom> columns) The important distinction wrt build() is that we may be building for a row that ends up being compacted away entirely, i.e., the input consists only of expired tombstones (or columns shadowed by expired tombstone).
void	finishAddingAtoms()
void	maybeWriteEmptyRowHeader()
RangeTombstone.Tracker	tombstoneTracker()
int	writtenAtomCount()

Methods inherited from class java.lang.Object

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

Constructor Detail

Builder

public Builder(ColumnFamily cf,
               java.nio.ByteBuffer key,
               DataOutputPlus output)

Builder

public Builder(ColumnFamily cf,
               java.nio.ByteBuffer key,
               DataOutputPlus output,
               OnDiskAtom.SerializerForWriting serializer)

Method Detail

tombstoneTracker

public RangeTombstone.Tracker tombstoneTracker()

writtenAtomCount

public int writtenAtomCount()

build

public ColumnIndex build(ColumnFamily cf)
                  throws java.io.IOException

Serializes the index into in-memory structure with all required components such as Bloom Filter, index block size, IndexInfo list

Parameters:

cf - Column family to create index for

Returns:

information about index - it's Bloom Filter, block size and IndexInfo list

Throws:

java.io.IOException

buildForCompaction

public ColumnIndex buildForCompaction(java.util.Iterator<OnDiskAtom> columns)
                               throws java.io.IOException

The important distinction wrt build() is that we may be building for a row that ends up being compacted away entirely, i.e., the input consists only of expired tombstones (or columns shadowed by expired tombstone). Thus, it is the caller's responsibility to decide whether to write the header for an empty row.

Throws:

java.io.IOException

add

public void add(OnDiskAtom column)
         throws java.io.IOException

Throws:

java.io.IOException

finishAddingAtoms

public void finishAddingAtoms()
                       throws java.io.IOException

Throws:

java.io.IOException

build

public ColumnIndex build()

maybeWriteEmptyRowHeader

public void maybeWriteEmptyRowHeader()
                              throws java.io.IOException

Throws:

java.io.IOException