ctypesio (version 0.1.2)

read_f64: Read floating point values from a connection

Description

Read floating point numbers into a standard R vector of doubles

Usage

read_f64(con, n = 1, endian = NULL)
read_f32(con, n = 1, endian = NULL)
read_f16(con, n = 1, endian = NULL)
read_bfloat(con, n = 1, endian = NULL)
read_dbl(con, n = 1, endian = NULL)
read_float(con, n = 1, endian = NULL)
read_half(con, n = 1, endian = NULL)

Value

vector of double precision floating point numbers

Arguments

con

Connection object or raw vector. Connection objects can be created with file(), url(), rawConnection() or any of the other many connection creation functions.

n

Number of elements to read. Default: 1

endian

Ordering of bytes within the file when reading multi-byte values. Possible values: 'big' or 'little'. Default: NULL indicates that endian option should be retrieved from the connection object if possible (where the user has used set_endian()) or otherwise will be set to "little"

Details

double precision

8 byte floating point numbers. read_f64() also available as read_dbl()

single precision

4 byte floating point numbers. read_f32() also available as read_float()

half precision

2 byte floating point numbers. read_f16() also available as read_half(). Consists of 1 sign bit, 5 bits for exponent and 10 bits for fraction.

bfloat

2 byte floating point numbers in the bfloat format read_bfloat(). Consits of 1 sign bit, 8 bits fo exponent and 7 bits for fraction.

See Also

Other data input functions: read_hex(), read_raw(), read_str(), read_uint8(), scan_dbl()

Examples

# Raw vector with 16 bytes (128 bits) of dummy data
data <- as.raw(1:16)
con <- rawConnection(data, 'rb')
read_f64(con, n = 1) # Read a 64-bit double-precision number
read_f16(con, n = 4) # Read 4 x 16-bit half-precision number
close(con)