org.apache.cassandra.db

Class DataRange

java.lang.Object

org.apache.cassandra.db.DataRange

Direct Known Subclasses:

DataRange.Paging

public class DataRange
extends java.lang.Object

Groups key range and column filter for range queries. The main "trick" of this class is that the column filter can only be obtained by providing the row key on which the column filter will be applied (which we always know before actually querying the columns). This allows the paging DataRange to return a filter for most rows but a potentially different ones for the starting and stopping key. Could allow more fancy stuff in the future too, like column filters that depend on the actual key value :)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DataRange.Paging

Field Summary

Fields

Modifier and Type	Field and Description
protected IDiskAtomFilter	columnFilter
protected boolean	selectFullRow

Constructor Summary

Constructors

Constructor and Description
DataRange(AbstractBounds<RowPosition> range, IDiskAtomFilter columnFilter)

Method Summary

Methods

Modifier and Type	Method and Description
static DataRange	allData(IPartitioner partitioner)
IDiskAtomFilter	columnFilter(java.nio.ByteBuffer rowKey) Returns a column filter that should be used for a particular row key.
boolean	contains(RowPosition pos)
static DataRange	forKeyRange(Range<Token> keyRange)
int	getLiveCount(ColumnFamily data, long now)
boolean	ignoredTombstonedPartitions() Returns true if tombstoned partitions should not be included in results or count towards the limit.
static boolean	isFullRowSlice(SliceQueryFilter filter)
boolean	isWrapAround()
AbstractBounds<RowPosition>	keyRange()
boolean	selectsFullRowFor(java.nio.ByteBuffer rowKey)
RowPosition	startKey()
RowPosition	stopKey()
void	updateColumnsLimit(int count) Sets a new limit on the number of (grouped) cells to fetch.

Methods inherited from class java.lang.Object

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

Field Detail

columnFilter

protected IDiskAtomFilter columnFilter

selectFullRow

protected final boolean selectFullRow

Constructor Detail

DataRange

public DataRange(AbstractBounds<RowPosition> range,
         IDiskAtomFilter columnFilter)

Method Detail

isFullRowSlice

public static boolean isFullRowSlice(SliceQueryFilter filter)

allData

public static DataRange allData(IPartitioner partitioner)

forKeyRange

public static DataRange forKeyRange(Range<Token> keyRange)

keyRange

public AbstractBounds<RowPosition> keyRange()

startKey

public RowPosition startKey()

stopKey

public RowPosition stopKey()

ignoredTombstonedPartitions

public boolean ignoredTombstonedPartitions()

Returns true if tombstoned partitions should not be included in results or count towards the limit. See CASSANDRA-8490 for more details on why this is needed (and done this way).

isWrapAround

public boolean isWrapAround()

contains

public boolean contains(RowPosition pos)

getLiveCount

public int getLiveCount(ColumnFamily data,
               long now)

selectsFullRowFor

public boolean selectsFullRowFor(java.nio.ByteBuffer rowKey)

columnFilter

public IDiskAtomFilter columnFilter(java.nio.ByteBuffer rowKey)

Returns a column filter that should be used for a particular row key. Note that in the case of paging, slice starts and ends may change depending on the row key.

updateColumnsLimit

public void updateColumnsLimit(int count)

Sets a new limit on the number of (grouped) cells to fetch. This is currently only used when the query limit applies to CQL3 rows.