org.apache.fury.format.row.binary

Class BinaryRow

java.lang.Object

org.apache.fury.format.row.binary.BinaryRow

All Implemented Interfaces:

Getters, Row, Setters

public class BinaryRow
extends Object
implements Row

A binary implementation of Row backed by binary buffer instead of java objects.

• Validity Bit Set Bitmap Region (1 bit/field) for tracking null values. Primitive type is always considered to be not null. Set bit to 1 indicate the value is not null, Set bit to 0 indicate null
• Fixed-Length 8-byte Values Region. if field isn't aligned, read any length-gt-1 value may need read multi times cache.
• Variable-Length Data Section

Equality comparison and hashing of rows can be performed on raw bytes since if two rows are identical so should be their bit-wise representation.

BinaryRow is inspired by Apache Spark tungsten UnsafeRow, the differences are
• Use arrow schema to describe meta.
• String support latin/utf16/utf8 encoding.
• Decimal use arrow decimal format.
• Variable-size field can be inline in fixed-size region if small enough.
• Allow skip padding by generate Row using aot to put offsets in generated code.
• The implementation support java/C++/python/etc..
• Support adding fields without breaking compatibility

Constructor Summary

Constructors

Constructor and Description
BinaryRow(org.apache.arrow.vector.types.pojo.Schema schema)

Method Summary

Modifier and Type	Method and Description
boolean	anyNull()
void	assertIndexIsValid(int index)
Row	copy()
BinaryArray	getArray(int ordinal)
int	getBaseOffset()
byte[]	getBinary(int ordinal)
boolean	getBoolean(int ordinal)
MemoryBuffer	getBuffer()
MemoryBuffer	getBuffer(int ordinal)
byte	getByte(int ordinal)
int	getDate(int ordinal)
BigDecimal	getDecimal(int ordinal)
float	getFloat32(int ordinal)
double	getFloat64(int ordinal)
short	getInt16(int ordinal)
int	getInt32(int ordinal)
long	getInt64(int ordinal)
BinaryMap	getMap(int ordinal)
int	getOffset(int ordinal)
org.apache.arrow.vector.types.pojo.Schema	getSchema()
int	getSizeInBytes()
String	getString(int ordinal)
BinaryRow	getStruct(int ordinal)
long	getTimestamp(int ordinal)
static boolean	isFixedLength(org.apache.arrow.vector.types.pojo.ArrowType type) If it is a fixed-length field, we can call this BinaryRow's setXX method for in-place updates.
boolean	isNullAt(int ordinal)
int	numFields()
void	pointTo(MemoryBuffer buffer, int offset, int sizeInBytes)
void	setBoolean(int ordinal, boolean value)
void	setByte(int ordinal, byte value)
void	setDate(int ordinal, int value)
void	setFloat32(int ordinal, float value)
void	setFloat64(int ordinal, double value)
void	setInt16(int ordinal, short value)
void	setInt32(int ordinal, int value)
void	setInt64(int ordinal, long value)
void	setNotNullAt(int ordinal)
void	setNullAt(int ordinal)
void	setTimestamp(int ordinal, long value)
byte[]	toBytes()
String	toDebugString()
Map<String,Object>	toMap()
String	toString()

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.fury.format.row.Getters

get, getBinary, getBoolean, getBuffer, getByte, getDate, getFloat32, getFloat64, getInt16, getInt32, getInt64, getString, getTimestamp

Methods inherited from interface org.apache.fury.format.row.Setters

setArray, setBinary, setBoolean, setByte, setDate, setDecimal, setFloat32, setFloat64, setInt16, setInt32, setInt64, setMap, setString, setStruct, setTimestamp

Constructor Detail

BinaryRow

public BinaryRow(org.apache.arrow.vector.types.pojo.Schema schema)

Method Detail

pointTo

public void pointTo(MemoryBuffer buffer,
                    int offset,
                    int sizeInBytes)

getSchema

public org.apache.arrow.vector.types.pojo.Schema getSchema()

Specified by:

getSchema in interface Row

numFields

public int numFields()

Specified by:

numFields in interface Row

getSizeInBytes

public int getSizeInBytes()

getBaseOffset

public int getBaseOffset()

getBuffer

public MemoryBuffer getBuffer()

getOffset

public int getOffset(int ordinal)

assertIndexIsValid

public void assertIndexIsValid(int index)

isNullAt

public boolean isNullAt(int ordinal)

Specified by:

isNullAt in interface Getters

anyNull

public boolean anyNull()

Specified by:

anyNull in interface Row

setNullAt

public void setNullAt(int ordinal)

Specified by:

setNullAt in interface Setters

setNotNullAt

public void setNotNullAt(int ordinal)

getDecimal

public BigDecimal getDecimal(int ordinal)

Specified by:

getDecimal in interface Getters

getStruct

public BinaryRow getStruct(int ordinal)

Specified by:

getStruct in interface Getters

getArray

public BinaryArray getArray(int ordinal)

Specified by:

getArray in interface Getters

getMap

public BinaryMap getMap(int ordinal)

Specified by:

getMap in interface Getters

copy

public Row copy()

Specified by:

copy in interface Row

toString

public String toString()

Overrides:

toString in class Object

toDebugString

public String toDebugString()

toMap

public Map<String,Object> toMap()

toBytes

public byte[] toBytes()

isFixedLength

public static boolean isFixedLength(org.apache.arrow.vector.types.pojo.ArrowType type)

If it is a fixed-length field, we can call this BinaryRow's setXX method for in-place updates. If it is variable-length field, can't use this method, because the underlying data is stored continuously.

getBuffer

public MemoryBuffer getBuffer(int ordinal)

Specified by:

getBuffer in interface Getters

getBoolean

public boolean getBoolean(int ordinal)

Specified by:

getBoolean in interface Getters

getByte

public byte getByte(int ordinal)

Specified by:

getByte in interface Getters

getInt16

public short getInt16(int ordinal)

Specified by:

getInt16 in interface Getters

getInt32

public int getInt32(int ordinal)

Specified by:

getInt32 in interface Getters

getInt64

public long getInt64(int ordinal)

Specified by:

getInt64 in interface Getters

getFloat32

public float getFloat32(int ordinal)

Specified by:

getFloat32 in interface Getters

getFloat64

public double getFloat64(int ordinal)

Specified by:

getFloat64 in interface Getters

getDate

public int getDate(int ordinal)

Specified by:

getDate in interface Getters

getTimestamp

public long getTimestamp(int ordinal)

Specified by:

getTimestamp in interface Getters

getString

public String getString(int ordinal)

Specified by:

getString in interface Getters

getBinary

public byte[] getBinary(int ordinal)

Specified by:

getBinary in interface Getters

setBoolean

public void setBoolean(int ordinal,
                       boolean value)

Specified by:

setBoolean in interface Setters

setByte

public void setByte(int ordinal,
                    byte value)

Specified by:

setByte in interface Setters

setInt16

public void setInt16(int ordinal,
                     short value)

Specified by:

setInt16 in interface Setters

setInt32

public void setInt32(int ordinal,
                     int value)

Specified by:

setInt32 in interface Setters

setInt64

public void setInt64(int ordinal,
                     long value)

Specified by:

setInt64 in interface Setters

setFloat32

public void setFloat32(int ordinal,
                       float value)

Specified by:

setFloat32 in interface Setters

setFloat64

public void setFloat64(int ordinal,
                       double value)

Specified by:

setFloat64 in interface Setters

setDate

public void setDate(int ordinal,
                    int value)

Specified by:

setDate in interface Setters

setTimestamp

public void setTimestamp(int ordinal,
                         long value)

Specified by:

setTimestamp in interface Setters