iceberg::RoaringPositionBitmap Class Reference

A bitmap that supports positive 64-bit positions, optimized for cases where most positions fit in 32 bits. More...

#include <roaring_position_bitmap.h>

Classes
struct	Impl

Public Member Functions
	RoaringPositionBitmap (RoaringPositionBitmap &&other) noexcept
RoaringPositionBitmap &	operator= (RoaringPositionBitmap &&other) noexcept
	RoaringPositionBitmap (const RoaringPositionBitmap &other)
RoaringPositionBitmap &	operator= (const RoaringPositionBitmap &other)
void	Add (int64_t pos)
	Sets a position in the bitmap.
void	AddRange (int64_t pos_start, int64_t pos_end)
	Sets a range of positions [pos_start, pos_end).
bool	Contains (int64_t pos) const
	Checks if a position is set in the bitmap.
bool	IsEmpty () const
	Returns true if the bitmap has no positions set.
size_t	Cardinality () const
	Returns the number of set positions in the bitmap.
void	Or (const RoaringPositionBitmap &other)
	Merges all positions from the other bitmap into this one (in-place union).
bool	Optimize ()
	Optimizes the bitmap by applying run-length encoding to containers where it is more space efficient than array or bitset representations.
void	ForEach (const std::function< void(int64_t)> &fn) const
	Iterates over all set positions in ascending order.
size_t	SerializedSizeInBytes () const
	Returns the serialized size in bytes.
Result< std::string >	Serialize () const
	Serializes using the portable format (little-endian).

Static Public Member Functions
static Result< RoaringPositionBitmap >	Deserialize (std::string_view bytes)
	Deserializes a bitmap from bytes.

Static Public Attributes
static constexpr int64_t	kMaxPosition = 0x7FFFFFFE80000000LL
	Maximum supported position (aligned with the Java implementation).

Detailed Description

A bitmap that supports positive 64-bit positions, optimized for cases where most positions fit in 32 bits.

Incoming 64-bit positions are divided into a 32-bit "key" using the most significant 4 bytes and a 32-bit position using the least significant 4 bytes. For each key, a 32-bit Roaring bitmap is maintained to store positions for that key.

Note

This class is used to represent deletion vectors. The Puffin reader/writer handle adding the additional required framing (length prefix, magic bytes, CRC-32) for deletion-vector-v1 persistence.

Member Function Documentation

Add()

void iceberg::RoaringPositionBitmap::Add

(

int64_t

pos

)

Sets a position in the bitmap.

Parameters

pos	the position (must be >= 0 and <= kMaxPosition)

Note

Invalid positions are silently ignored

AddRange()

void iceberg::RoaringPositionBitmap::AddRange	(	int64_t	pos_start,
		int64_t	pos_end
	)

Sets a range of positions [pos_start, pos_end).

Parameters

pos_start	the start of the range (inclusive), clamped to 0
pos_end	the end of the range (exclusive), clamped to kMaxPosition + 1

Note

If pos_start > pos_end, the call is silently ignored. If pos_start == pos_end, this method does nothing. Positions outside [0, kMaxPosition] are silently ignored.

Contains()

bool iceberg::RoaringPositionBitmap::Contains

(

int64_t

pos

)

const

Checks if a position is set in the bitmap.

Parameters

pos	the position to check

Returns

true if the position is set, false otherwise (including invalid positions)

Optimize()

bool iceberg::RoaringPositionBitmap::Optimize

(

)

Optimizes the bitmap by applying run-length encoding to containers where it is more space efficient than array or bitset representations.

Returns

true if any container was changed

---

The documentation for this class was generated from the following files:

• iceberg/deletes/roaring_position_bitmap.h
• iceberg/deletes/roaring_position_bitmap.cc