PLAIN-encoded Parquet column decoder. More...

#include <column_reader.hpp>

Public Member Functions
	ColumnReader (PhysicalType type, const uint8_t *data, size_t size, int64_t num_values, int32_t type_length=-1)
	Construct a reader over raw PLAIN-encoded page data.

expected< bool >	read_bool ()
	Read a single BOOLEAN value (bit-packed, LSB first).

expected< int32_t >	read_int32 ()
	Read a single INT32 value (4 bytes little-endian).

expected< int64_t >	read_int64 ()
	Read a single INT64 value (8 bytes little-endian).

expected< float >	read_float ()
	Read a single FLOAT value (4 bytes little-endian, IEEE 754).

expected< double >	read_double ()
	Read a single DOUBLE value (8 bytes little-endian, IEEE 754).

expected< std::string >	read_string ()
	Read a single BYTE_ARRAY value as a std::string.

expected< std::string_view >	read_string_view ()
	Read a single BYTE_ARRAY value as a non-owning std::string_view.

expected< std::vector< uint8_t > >	read_bytes ()
	Read a single BYTE_ARRAY or FIXED_LEN_BYTE_ARRAY value as raw bytes.

expected< void >	read_batch_bool (bool *out, size_t count)
	Read a batch of BOOLEAN values into `out`.

expected< void >	read_batch_int32 (int32_t *out, size_t count)
	Read a batch of INT32 values via bulk memcpy.

expected< void >	read_batch_int64 (int64_t *out, size_t count)
	Read a batch of INT64 values via bulk memcpy.

expected< void >	read_batch_float (float *out, size_t count)
	Read a batch of FLOAT values via bulk memcpy.

expected< void >	read_batch_double (double *out, size_t count)
	Read a batch of DOUBLE values via bulk memcpy.

expected< void >	read_batch_string (std::string *out, size_t count)
	Read a batch of BYTE_ARRAY values as strings.

template<typename T >
expected< T >	read ()
	Read a single value of type `T`, dispatching to the correct typed reader.

template<typename T >
expected< void >	read_batch (T *out, size_t count)
	Read a batch of `count` values of type `T` into `out`.

int64_t	values_remaining () const
	Number of values not yet read from this page.

bool	has_next () const
	Whether there is at least one more value to read.

PhysicalType	type () const
	The Parquet physical type of this column.

size_t	position () const
	Current byte offset within the page data buffer.

Detailed Description

PLAIN-encoded Parquet column decoder.

Wraps a raw data page buffer and decodes values one at a time or in batches. The reader maintains a cursor position and a count of values read, returning an error on type mismatch, buffer overrun, or exhaustion.

Note: ColumnReader does not own the data buffer. The caller must ensure the buffer remains valid for the reader's lifetime.

See also: ColumnWriter (the encoding counterpart); MmapParquetReader (constructs ColumnReaders over mmap'd pages)

Definition at line 46 of file column_reader.hpp.

Constructor & Destructor Documentation

◆ ColumnReader()

signet::forge::ColumnReader::ColumnReader	(	PhysicalType	type,
		const uint8_t *	data,
		size_t	size,
		int64_t	num_values,
		int32_t	type_length = `-1`
	)

inline

Construct a reader over raw PLAIN-encoded page data.

Parameters

type	The physical type of the column.
data	Pointer to the start of the page data buffer.
size	Size of the page data in bytes.
num_values	Number of values encoded in this page.
type_length	For FIXED_LEN_BYTE_ARRAY columns, the fixed byte length per value (ignored for other types).

Definition at line 56 of file column_reader.hpp.

Member Function Documentation

◆ has_next()

bool signet::forge::ColumnReader::has_next ( ) const

inline

Whether there is at least one more value to read.

Definition at line 566 of file column_reader.hpp.

◆ position()

size_t signet::forge::ColumnReader::position ( ) const

inline

Current byte offset within the page data buffer.

Definition at line 576 of file column_reader.hpp.

◆ read()

template<typename T >

expected< T > signet::forge::ColumnReader::read ( )

inline

Read a single value of type T, dispatching to the correct typed reader.

Supported types: bool, int32_t, int64_t, float, double, std::string, std::string_view, std::vector<uint8_t>.

Template Parameters

T	The value type to decode.

Returns: The decoded value, or an error.

Definition at line 504 of file column_reader.hpp.

◆ read_batch()

template<typename T >

expected< void > signet::forge::ColumnReader::read_batch	(	T *	out,
		size_t	count
	)

inline

Read a batch of count values of type T into out.

Dispatches to the correct typed batch reader. Supported types: bool, int32_t, int64_t, float, double, std::string.

Template Parameters

T	The value type to decode.

Parameters

out	Pre-allocated buffer of at least `count` elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 537 of file column_reader.hpp.

◆ read_batch_bool()

expected< void > signet::forge::ColumnReader::read_batch_bool	(	bool *	out,
		size_t	count
	)

inline

Read a batch of BOOLEAN values into out.

Parameters

out	Pre-allocated buffer of at least `count` elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 319 of file column_reader.hpp.

◆ read_batch_double()

expected< void > signet::forge::ColumnReader::read_batch_double	(	double *	out,
		size_t	count
	)

inline

Read a batch of DOUBLE values via bulk memcpy.

Parameters

out	Pre-allocated buffer of at least `count` elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 433 of file column_reader.hpp.

◆ read_batch_float()

expected< void > signet::forge::ColumnReader::read_batch_float	(	float *	out,
		size_t	count
	)

inline

Read a batch of FLOAT values via bulk memcpy.

Parameters

out	Pre-allocated buffer of at least `count` elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 405 of file column_reader.hpp.

◆ read_batch_int32()

expected< void > signet::forge::ColumnReader::read_batch_int32	(	int32_t *	out,
		size_t	count
	)

inline

Read a batch of INT32 values via bulk memcpy.

Parameters

out	Pre-allocated buffer of at least `count` elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 349 of file column_reader.hpp.

◆ read_batch_int64()

expected< void > signet::forge::ColumnReader::read_batch_int64	(	int64_t *	out,
		size_t	count
	)

inline

Read a batch of INT64 values via bulk memcpy.

Parameters

out	Pre-allocated buffer of at least `count` elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 377 of file column_reader.hpp.

◆ read_batch_string()

expected< void > signet::forge::ColumnReader::read_batch_string	(	std::string *	out,
		size_t	count
	)

inline

Read a batch of BYTE_ARRAY values as strings.

Parameters

out	Pre-allocated buffer of at least `count` std::string elements.
count	Number of values to read.

Returns: Void on success, or an error.

Definition at line 461 of file column_reader.hpp.

◆ read_bool()

expected< bool > signet::forge::ColumnReader::read_bool ( )

inline

Read a single BOOLEAN value (bit-packed, LSB first).

Returns: The decoded boolean, or an error on type mismatch / exhaustion.

Definition at line 76 of file column_reader.hpp.

◆ read_bytes()

expected< std::vector< uint8_t > > signet::forge::ColumnReader::read_bytes ( )

inline

Read a single BYTE_ARRAY or FIXED_LEN_BYTE_ARRAY value as raw bytes.

For BYTE_ARRAY, reads a 4-byte LE length prefix then the payload. For FIXED_LEN_BYTE_ARRAY, reads exactly type_length bytes.

Returns: A vector of the raw bytes, or an error.

Definition at line 264 of file column_reader.hpp.

◆ read_double()

expected< double > signet::forge::ColumnReader::read_double ( )

inline

Read a single DOUBLE value (8 bytes little-endian, IEEE 754).

Returns: The decoded double, or an error on type mismatch / exhaustion.

Definition at line 167 of file column_reader.hpp.

◆ read_float()

expected< float > signet::forge::ColumnReader::read_float ( )

inline

Read a single FLOAT value (4 bytes little-endian, IEEE 754).

Returns: The decoded float, or an error on type mismatch / exhaustion.

Definition at line 145 of file column_reader.hpp.

◆ read_int32()

expected< int32_t > signet::forge::ColumnReader::read_int32 ( )

inline

Read a single INT32 value (4 bytes little-endian).

Returns: The decoded int32_t, or an error on type mismatch / exhaustion.

Definition at line 101 of file column_reader.hpp.

◆ read_int64()

expected< int64_t > signet::forge::ColumnReader::read_int64 ( )

inline

Read a single INT64 value (8 bytes little-endian).

Returns: The decoded int64_t, or an error on type mismatch / exhaustion.

Definition at line 123 of file column_reader.hpp.

◆ read_string()

expected< std::string > signet::forge::ColumnReader::read_string ( )

inline

Read a single BYTE_ARRAY value as a std::string.

PLAIN encoding: 4-byte LE length prefix followed by raw bytes.

Returns: The decoded string, or an error on type mismatch / exhaustion.

Definition at line 192 of file column_reader.hpp.

◆ read_string_view()

expected< std::string_view > signet::forge::ColumnReader::read_string_view ( )

inline

Read a single BYTE_ARRAY value as a non-owning std::string_view.

The returned view points directly into the page data buffer, so it is only valid as long as the underlying buffer is alive.

Returns: A string_view into the page data, or an error.

Definition at line 228 of file column_reader.hpp.

◆ type()

PhysicalType signet::forge::ColumnReader::type ( ) const

inline

The Parquet physical type of this column.

Definition at line 571 of file column_reader.hpp.

◆ values_remaining()

int64_t signet::forge::ColumnReader::values_remaining ( ) const

inline

Number of values not yet read from this page.

Definition at line 561 of file column_reader.hpp.

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

include/signet/column_reader.hpp

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ ColumnReader()

Member Function Documentation

◆ has_next()

◆ position()

◆ read()

◆ read_batch()

◆ read_batch_bool()

◆ read_batch_double()

◆ read_batch_float()

◆ read_batch_int32()

◆ read_batch_int64()

◆ read_batch_string()

◆ read_bool()

◆ read_bytes()

◆ read_double()

◆ read_float()

◆ read_int32()

◆ read_int64()

◆ read_string()

◆ read_string_view()

◆ type()

◆ values_remaining()