org.apache.cassandra.db

Class AtomicSortedColumns

java.lang.Object

org.apache.cassandra.db.ColumnFamily

org.apache.cassandra.db.AtomicSortedColumns

All Implemented Interfaces:

java.lang.Iterable<Column>, IRowCacheEntry

public class AtomicSortedColumns
extends ColumnFamily

A thread-safe and atomic ISortedColumns implementation. Operations (in particular addAll) on this implemenation are atomic and isolated (in the sense of ACID). Typically a addAll is guaranteed that no other thread can see the state where only parts but not all columns have been added. The implementation uses snaptree (https://github.com/nbronson/snaptree), and in particular it's copy-on-write clone operation to achieve its atomicity guarantee. WARNING: removing element through getSortedColumns().iterator() is *not* isolated of other operations and could actually be fully ignored in the face of a concurrent. Don't use it unless in a non-concurrent context.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.cassandra.db.ColumnFamily

ColumnFamily.Factory<T extends ColumnFamily>

Field Summary

Fields

Modifier and Type	Field and Description
static ColumnFamily.Factory<AtomicSortedColumns>	factory

Fields inherited from class org.apache.cassandra.db.ColumnFamily

metadata, serializer

Method Summary

Methods

Modifier and Type	Method and Description
void	addAll(ColumnFamily cm, Allocator allocator, com.google.common.base.Function<Column,Column> transformation) Adds all the columns of a given column map to this column map.
Pair<java.lang.Long,java.lang.Long>	addAllWithSizeDelta(ColumnFamily cm, Allocator allocator, com.google.common.base.Function<Column,Column> transformation, SecondaryIndexManager.Updater indexer) This is only called by Memtable.resolve, so only AtomicSortedColumns needs to implement it.
void	addColumn(Column column, Allocator allocator) Adds a column to this column map.
void	clear() Clear this column family, removing all columns and deletion info.
ColumnFamily	cloneMe() Clones the column map.
void	delete(DeletionInfo info)
void	delete(DeletionTime delTime)
protected void	delete(RangeTombstone tombstone)
DeletionInfo	deletionInfo()
Column	getColumn(java.nio.ByteBuffer name) Get a column given its name, returning null if the column is not present.
int	getColumnCount() Returns the number of columns in this map.
java.util.SortedSet<java.nio.ByteBuffer>	getColumnNames() Returns an iterable with the names of columns in this column map in the same order as the underlying columns themselves.
AbstractType<?>	getComparator()
ColumnFamily.Factory	getFactory() Returns the factory used for this ISortedColumns implementation.
java.util.Collection<Column>	getReverseSortedColumns() Returns the columns of this column map as a collection.
java.util.Collection<Column>	getSortedColumns() Returns the columns of this column map as a collection.
boolean	isInsertReversed() Returns if this map only support inserts in reverse order.
java.util.Iterator<Column>	iterator(ColumnSlice[] slices) Returns an iterator over the columns of this map that returns only the matching @param slices.
void	purgeTombstones(int gcBefore) Purges top-level and range tombstones whose localDeletionTime is older than gcBefore.
boolean	replace(Column oldColumn, Column newColumn) Replace oldColumn if present by newColumn.
java.util.Iterator<Column>	reverseIterator(ColumnSlice[] slices) Returns a reversed iterator over the columns of this map that returns only the matching @param slices.
void	setDeletionInfo(DeletionInfo newInfo)

Methods inherited from class org.apache.cassandra.db.ColumnFamily

addAll, addAtom, addColumn, addColumn, addColumn, addCounter, addIfRelevant, addTombstone, addTombstone, asMap, batchRemoveIterator, cloneMeShallow, cloneMeShallow, cloneMeShallow, dataSize, delete, diff, diff, digest, equals, fromBytes, getColumnStats, getType, hashCode, hasIrrelevantData, hasOnlyTombstones, id, inOrderDeletionTester, isEmpty, isMarkedForDelete, iterator, maxTimestamp, metadata, resolve, resolve, reverseIterator, toBytes, toString, updateDigest

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Field Detail

factory

public static final ColumnFamily.Factory<AtomicSortedColumns> factory

Method Detail

getComparator

public AbstractType<?> getComparator()

Overrides:

getComparator in class ColumnFamily

Returns:

the comparator whose sorting order the contained columns conform to

getFactory

public ColumnFamily.Factory getFactory()

Description copied from class: ColumnFamily

Returns the factory used for this ISortedColumns implementation.

Specified by:

getFactory in class ColumnFamily

cloneMe

public ColumnFamily cloneMe()

Description copied from class: ColumnFamily

Clones the column map.

Specified by:

cloneMe in class ColumnFamily

deletionInfo

public DeletionInfo deletionInfo()

Specified by:

deletionInfo in class ColumnFamily

delete

public void delete(DeletionTime delTime)

Specified by:

delete in class ColumnFamily

delete

protected void delete(RangeTombstone tombstone)

Specified by:

delete in class ColumnFamily

delete

public void delete(DeletionInfo info)

Specified by:

delete in class ColumnFamily

setDeletionInfo

public void setDeletionInfo(DeletionInfo newInfo)

Specified by:

setDeletionInfo in class ColumnFamily

purgeTombstones

public void purgeTombstones(int gcBefore)

Description copied from class: ColumnFamily

Purges top-level and range tombstones whose localDeletionTime is older than gcBefore.

Specified by:

purgeTombstones in class ColumnFamily

Parameters:

gcBefore - a timestamp (in seconds) before which tombstones should be purged

addColumn

public void addColumn(Column column,
             Allocator allocator)

Description copied from class: ColumnFamily

Adds a column to this column map. If a column with the same name is already present in the map, it will be replaced by the newly added column.

Specified by:

addColumn in class ColumnFamily

addAll

public void addAll(ColumnFamily cm,
          Allocator allocator,
          com.google.common.base.Function<Column,Column> transformation)

Description copied from class: ColumnFamily

Adds all the columns of a given column map to this column map. This is equivalent to: for (Column c : cm) addColumn(c, ...); but is potentially faster.

Specified by:

addAll in class ColumnFamily

addAllWithSizeDelta

public Pair<java.lang.Long,java.lang.Long> addAllWithSizeDelta(ColumnFamily cm,
                                                      Allocator allocator,
                                                      com.google.common.base.Function<Column,Column> transformation,
                                                      SecondaryIndexManager.Updater indexer)

This is only called by Memtable.resolve, so only AtomicSortedColumns needs to implement it.

Returns:

the difference in size seen after merging the given columns and minimum time delta between timestamps of two reconciled columns.

replace

public boolean replace(Column oldColumn,
              Column newColumn)

Description copied from class: ColumnFamily

Replace oldColumn if present by newColumn. Returns true if oldColumn was present and thus replaced. oldColumn and newColumn should have the same name.

Specified by:

replace in class ColumnFamily

clear

public void clear()

Description copied from class: ColumnFamily

Clear this column family, removing all columns and deletion info.

Specified by:

clear in class ColumnFamily

getColumn

public Column getColumn(java.nio.ByteBuffer name)

Description copied from class: ColumnFamily

Get a column given its name, returning null if the column is not present.

Specified by:

getColumn in class ColumnFamily

getColumnNames

public java.util.SortedSet<java.nio.ByteBuffer> getColumnNames()

Description copied from class: ColumnFamily

Returns an iterable with the names of columns in this column map in the same order as the underlying columns themselves.

Specified by:

getColumnNames in class ColumnFamily

getSortedColumns

public java.util.Collection<Column> getSortedColumns()

Description copied from class: ColumnFamily

Returns the columns of this column map as a collection. The columns in the returned collection should be sorted as the columns in this map.

Specified by:

getSortedColumns in class ColumnFamily

getReverseSortedColumns

public java.util.Collection<Column> getReverseSortedColumns()

Description copied from class: ColumnFamily

Returns the columns of this column map as a collection. The columns in the returned collection should be sorted in reverse order of the columns in this map.

Specified by:

getReverseSortedColumns in class ColumnFamily

getColumnCount

public int getColumnCount()

Description copied from class: ColumnFamily

Returns the number of columns in this map.

Specified by:

getColumnCount in class ColumnFamily

iterator

public java.util.Iterator<Column> iterator(ColumnSlice[] slices)

Description copied from class: ColumnFamily

Returns an iterator over the columns of this map that returns only the matching @param slices. The provided slices must be in order and must be non-overlapping.

Specified by:

iterator in class ColumnFamily

reverseIterator

public java.util.Iterator<Column> reverseIterator(ColumnSlice[] slices)

Description copied from class: ColumnFamily

Returns a reversed iterator over the columns of this map that returns only the matching @param slices. The provided slices must be in reversed order and must be non-overlapping.

Specified by:

reverseIterator in class ColumnFamily

isInsertReversed

public boolean isInsertReversed()

Description copied from class: ColumnFamily

Returns if this map only support inserts in reverse order.

Specified by:

isInsertReversed in class ColumnFamily