Convert byte array to float array in Python 3

Question

I am sending 1.0 and 2.0 as binary from C program to STDOUT:

    float array[2] = { 1.0, 2.0 };
    fwrite(&array[0], sizeof(array[0]), sizeof(array) / sizeof(array[0]), stdout);
    fflush(stdout);

Python program reads STDOUT as:

b'\x00\x00\x80?\x00\x00\x00@'

When I try to convert this back to float, this works:

struct.unpack('<f', array[0:4]) # 1.0
struct.unpack('<f', array[4:8]) # 2.0

However, if I try to decode the whole array at once, it fails:

struct.unpack('<f', array)

with the following message:

error: unpack requires a buffer of 4 bytes

Is it possible to decode the whole array at once, or I should decode in a loop each float separately?

Theobaldus · Accepted Answer · 2020-10-19 12:57:44Z

5

Using struct

Base on the struct module documentation the struct.unpack() function must have defined exact format for unpacking multiple values at once. So if you need to use struct module, then you either have to define the format using format as '<ff' or iterate over the array using struct.iter_unpack(format, buffer).

Using array

Another option is to use array module, which allows you to decode the binary data at once. You can change endianness using array.byteswap() method. The only restriction is that the whole array must have same type.

import array

arr = array.array('f', b'\x00\x00\x80?\x00\x00\x00@')

arr.tolist()
# [1.0, 2.0]

# change endianness
arr.byteswap()

arr.tolist()
# [4.600602988224807e-41, 8.96831017167883e-44]

edited Oct 19, 2020 at 12:57

answered Oct 19, 2020 at 12:52

Theobaldus

1437 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

jsbueno · Accepted Answer · 2020-10-19 13:07:06Z

1

Python's strcut provide some essential tooling, but is a bit clumsy to work with. In particular, for this, it'd have to yield all your floats as a tuple of Python floats (which are "enormous" objetcs with 10s of bytes each), and then insert then into an array.

Fortunatelly, Python's array.array class itself can fill in its values given a bytes object, or even read then directly from a file.


import array

data = b'\x00\x00\x80?\x00\x00\x00@'
arr = array.array('f')
arr.frombytes(data)
print(arr)

Yields:

array('f', [1.0, 2.0])

(to read directly from a file, use the fromfile method)

answered Oct 19, 2020 at 13:07

jsbueno

113k11 gold badges159 silver badges239 bronze badges

Collectives™ on Stack Overflow

Convert byte array to float array in Python 3

2 Answers 2

Using struct

Using array

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Using struct

Using array

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related