org.knime.core.data.vector.bitvector

Class SparseBitVectorCell

java.lang.Object

org.knime.core.data.DataCell

org.knime.core.data.vector.bitvector.SparseBitVectorCell

All Implemented Interfaces:

Serializable, DataValue, BitVectorValue

public class SparseBitVectorCell
extends DataCell
implements BitVectorValue

Stores Zeros and Ones in a vector, i.e. with fixed positions. The vector has a fixed length.
Implementation assumes that the vector is only sparsely populated with '1's. It stores the indices of the ones. For densely populated vectors DenseBitVector is more suitable.
The length of the vector is restricted to Long.MAX_VALUE (i.e. 9223372036854775807). The number of ones that can be stored is limited to Integer.MAX_VALUE (which is 2147483647), in which case it uses about 16Gbyte of memory.

Author:

ohl, University of Konstanz

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SparseBitVectorCell.Factory Factory for SparseBitVectorCells.
static class	SparseBitVectorCell.SparseBitVectorSerializer Factory for (de-)serializing a DenseBitVectorCell.

Nested classes/interfaces inherited from interface org.knime.core.data.vector.bitvector.BitVectorValue

BitVectorValue.BitVectorUtilityFactory

Nested classes/interfaces inherited from interface org.knime.core.data.DataValue

DataValue.UtilityFactory

Field Summary

Fields

Modifier and Type	Field and Description
static DataType	TYPE Convenience access member for DataType.getType(SparseBitVectorCell.class).

Fields inherited from interface org.knime.core.data.vector.bitvector.BitVectorValue

MAX_DISPLAY_BITS, UTILITY

Constructor Summary

Constructors

Constructor and Description
SparseBitVectorCell(SparseBitVector bitVector) Use the SparseBitVectorCellFactory to create instances of this cell.

Method Summary

Modifier and Type	Method and Description
long	cardinality() Returns the number of set bits (bits with value '1') in the vector.
(package private) long	cardinalityOfIntersection(SparseBitVectorCell bitVectorCell)
(package private) long	cardinalityOfRelativeComplement(SparseBitVectorCell bitVectorCell)
protected boolean	equalContent(DataValue otherValue) This method is called when two cell of different classes but with the same preferred value are compared (e.g.
protected boolean	equalsDataCell(DataCell dc) Derived classes implement their specific equals function here.
boolean	get(long index) Returns the value of the specified bit.
SparseBitVector	getBitVectorCopy() Returns a clone of the internal dense bit vector.
static DataCellSerializer<SparseBitVectorCell>	getCellSerializer() Deprecated. use DataTypeRegistry.getSerializer(Class) instead
int	hashCode() This method must be implemented in order to ensure that two equal DataCell objects return the same hash code.
boolean	isEmpty() Returns true, if all bits in the vector are cleared.
long	length() Returns the length of the bit vector.
long	nextClearBit(long startIdx) Finds the next bit not set (that is '0') on or after the specified index.
long	nextSetBit(long startIdx) Finds the next bit set to one on or after the specified index.
String	toBinaryString() Returns the binary string representation of the bits in this vector.
String	toHexString() Returns the hex representation of the bits in this vector.
String	toString() Returns the String representation of this cell's value.

Methods inherited from class org.knime.core.data.DataCell

equals, getType, isMissing

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.knime.core.data.vector.bitvector.BitVectorValue

equalContent

Field Detail

TYPE

public static final DataType TYPE

Convenience access member for DataType.getType(SparseBitVectorCell.class).

See Also:

DataType.getType(Class)

Constructor Detail

SparseBitVectorCell

SparseBitVectorCell(SparseBitVector bitVector)

Use the SparseBitVectorCellFactory to create instances of this cell.

Parameters:

bitVector - the bit vector to store in this cell.

Method Detail

getCellSerializer

@Deprecated
public static final DataCellSerializer<SparseBitVectorCell> getCellSerializer()

Deprecated. use DataTypeRegistry.getSerializer(Class) instead

Returns the factory to read/write DataCells of this class from/to a DataInput/DataOutput. This method is called via reflection.

Returns:

A serializer for reading/writing cells of this kind.

See Also:

DataCell

equalsDataCell

protected boolean equalsDataCell(DataCell dc)

Derived classes implement their specific equals function here. The argument is guaranteed to be not null or a missing value, to be of the same class like this.

Specified by:

equalsDataCell in class DataCell

Parameters:

dc - the cell to compare this to

Returns:

true if this is equal to the argument, false if not

equalContent

protected boolean equalContent(DataValue otherValue)

This method is called when two cell of different classes but with the same preferred value are compared (e.g. SdfCell and SdfAdapterCell or MolCell and MolBlobCell). All such cell classes should override this method with the same implementation, usually by comparing both preferred values. You can assume the this cell and the other cell have the same preferred value class, i.e. you can cast the argument to the preferred value.
The default implementation returns false.

Overrides:

equalContent in class DataCell

Parameters:

otherValue - the other data value

Returns:

true if the content of both cells is the same, false otherwise

hashCode

public int hashCode()

This method must be implemented in order to ensure that two equal DataCell objects return the same hash code. Note that two cells can be equal according to DataCell.equalsDataCell(DataCell) and DataCell.equalContent(DataValue) so two different cells implementations with the same preferred value must have the same hash code.

Specified by:

hashCode in class DataCell

Returns:

the hash code of your specific DataCell

See Also:

Object.hashCode(), DataCell.equals(java.lang.Object)

toString

public String toString()

Returns the String representation of this cell's value.

Specified by:

toString in class DataCell

Returns:

a String representation of this cell's value

toHexString

public String toHexString()

Returns the hex representation of the bits in this vector. Each character in the result represents 4 bits (with the characters '0' - '9' and 'A' - 'F'). The character at string position (length - 1) holds the lowest bits (bit 0 to 3), the character at position 0 represents the bits with the largest index in the vector. If the length of the vector is larger than (Integer.MAX_VALUE - 1) * 4 (i.e. 8589934584), the result is truncated (and ends with ...).

Specified by:

toHexString in interface BitVectorValue

Returns:

the hex representation of this bit vector.

toBinaryString

public String toBinaryString()

Returns the binary string representation of the bits in this vector. Each character in the result represents one bit - a '1' stands for a set bit, a '0' represents a cleared bit. The character at string position (length - 1) holds the bit with index 0, the character at position 0 represents the bits with the largest index in the vector. If the length of the vector is larger than (Integer.MAX_VALUE - 3) (i.e. 2147483644), the result is truncated (and ends with ...).

Specified by:

toBinaryString in interface BitVectorValue

Returns:

the binary (0/1) representation of this bit vector.

getBitVectorCopy

public SparseBitVector getBitVectorCopy()

Returns a clone of the internal dense bit vector.

Returns:

a copy of the internal dense bit vector.

cardinality

public long cardinality()

Returns the number of set bits (bits with value '1') in the vector.

Specified by:

cardinality in interface BitVectorValue

Returns:

the number of set bits (bits with value '1') in the vector.

get

public boolean get(long index)

Returns the value of the specified bit.

Specified by:

get in interface BitVectorValue

Parameters:

index - the index of the bit to test

Returns:

true if the bit at the specified index is set, false if it is zero.

isEmpty

public boolean isEmpty()

Returns true, if all bits in the vector are cleared.

Specified by:

isEmpty in interface BitVectorValue

Returns:

true, if all bits are zero, false, if at least one bit is set.

length

public long length()

Returns the length of the bit vector. The number of stored bits.

Specified by:

length in interface BitVectorValue

Returns:

the number of bits stored in the vector

nextClearBit

public long nextClearBit(long startIdx)

Finds the next bit not set (that is '0') on or after the specified index. Returns an index larger than or equal the provided index, or -1 if no bit is cleared after the startIdx. (It is okay to pass an index larger than the length of the vector.)

Specified by:

nextClearBit in interface BitVectorValue

Parameters:

startIdx - the first index to look for '0's.

Returns:

the index of the next cleared bit, which is on or after the provided startIdx. Or -1 if the vector contains no zero anymore.

nextSetBit

public long nextSetBit(long startIdx)

Finds the next bit set to one on or after the specified index. Returns an index larger than or equal the provided index, or -1 if no bit is set after the startIdx. (Is okay to pass an index larger than the length of the vector.)

Specified by:

nextSetBit in interface BitVectorValue

Parameters:

startIdx - the first index to look for '1's. (It is allowed to pass an index larger then the vector's length.)

Returns:

the index of the next bit set to one, which is on or after the provided startIdx.

cardinalityOfIntersection

long cardinalityOfIntersection(SparseBitVectorCell bitVectorCell)

Parameters:

bitVectorCell - the other cell containing the operand

Returns:

the cardinality of intersection

See Also:

BitVectorUtil.cardinalityOfIntersection(BitVectorValue, BitVectorValue)

cardinalityOfRelativeComplement

long cardinalityOfRelativeComplement(SparseBitVectorCell bitVectorCell)

Parameters:

bitVectorCell - the other cell containing the operand

Returns:

the cardinality of intersection

See Also:

BitVectorUtil.cardinalityOfRelativeComplement(BitVectorValue, BitVectorValue)