CODEWITHSUNDEEP
python-3.xencodingbase64
StanGeo
StanGeo

Reputation: 429

Convert float32 bytes to base64 encoded bytes

I have a float value 16.75 which needs to be encoded into base64 binary.

I am using the website https://cryptii.com/pipes/binary-to-base64 as a reference for validation. 16.75 encoded into 32bit floating point format results in 01000001 10000110 00000000 00000000 which gets a base64 encoding value of QYYAAA==

Using the accepted answer from Python, base64, float unfortunately leads me to a different answer still.

xx = base64.encodebytes(struct.pack('<d', 16.75))
print(xx) #b'AAAAAADAMEA=\n'
xx = struct.unpack('<d', base64.decodebytes(xx))
print(xx) #(16.75,)

Can someone guide me where am I going wrong? Do post your questions in comments - will try to answer as early as possible.

Upvotes: 1

Views: 870

Answers (1)

Marcel Wilson
Marcel Wilson

Reputation: 4572

The problem you're running into is the differences in byte order and size of float.

Python floats are doubles. The binary you're feeding into the Cryptii website is a float32 representation.

To match the answer you were looking up, you'd need to create the struct using "!f" (big-endian float32)

a = base64.encodebytes(struct.pack('!f', 16.75)) 
print(a) # b'QYYAAA==\n'

To get cryptii to return the answer you were creating in python with your example, you'd need to convert the float to binary using "<d" (little-endian double):

print(''.join('{:0>8b}'.format(c) for c in struct.pack('<d', 16.75)))

0000000000000000000000000000000000000000110000000011000001000000

Upvotes: 2

Related Questions