Converting binary representation to signed 64 bit integer in Python

Question

Against the advice of online resources, I'm using Python to do some simple bit shifting in one of my scripts. The bit shifting is fine, it's representing the result that's difficult. Let me explain.

I have a 64 bit binary representation

1000010101010111010101010101010101010101010101010111010101010101

which, when represented as a signed integer, should be negative. However, Python defaults to type long, preventing the number from being signed by the first bit, yielding the number 9608242155010487637.

How can I get Python to recognize the sign bit in 64 bits?

To clarify, I'm using Python 2.7.

You may find you answer here: stackoverflow.com/questions/5305456/… stackoverflow.com/questions/27016919/… — Loïc Poncin
– Loïc Poncin, Commented Feb 22, 2017 at 17:06
@LoïcPoncin It's great that you are looking for dupe targets, but neither of those questions are particularly relevant to this one. — PM 2Ring
– PM 2Ring, Commented Feb 22, 2017 at 17:15

wim · Accepted Answer · 2017-02-22 17:08:13Z

3

Try the bitstring module:

>>> from bitstring import BitArray
>>> s = '1000010101010111010101010101010101010101010101010111010101010101'
>>> BitArray(bin=s).int
-8838501918699063979

answered Feb 22, 2017 at 17:08

wim

368k114 gold badges681 silver badges816 bronze badges

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

Comments

Mark Reed · Accepted Answer · 2025-08-19 16:35:46Z

3

You can use struct, like this:

>>> import struct
>>> struct.unpack('>q',struct.pack('>Q',
int('1000010101010111010101010101010101010101010101010111010101010101',
2)))
#=> (-8838501918699063979,)

The result is a tuple, which you can then request the first element of:

>>> struct.unpack('>q',struct.pack('>Q',
int('1000010101010111010101010101010101010101010101010111010101010101',
2)))[0]
#=> -8838501918699063979

edited Aug 19 at 16:35

answered Feb 22, 2017 at 17:07

Mark Reed

96k17 gold badges149 silver badges189 bronze badges

3 Comments

DonutGaz Over a year ago

I'm getting a struct error struct.error: unpack requires a string argument of length 4

DonutGaz Over a year ago

I changed the l argument to q in order to represent the 64 bit signed long, which fixed this error and gave me the correct result.

Mark Reed Aug 19 at 16:36

@DonutGaz I used a platform-dependent format string, which I should not have done. Edited to a platform-independent version.

Markus-Hermann · Accepted Answer · 2025-08-19 11:21:41Z

It's been a looong time ... . Still, here is a basic-python-only solution to round it all off.
First, negative integers are just positive integers with an offset. Take a char, i.e., 8 bit.
Runs from (let's consider big endian, i.e., most significant bit (acting sign!) first, least significant bit last):

00000000 == 0
00000001 == 1
..
01111110 == 126
01111111 == 127

and then it flips around. All the way to

10000000 == -128
10000001 == -127
..
11111110 == -2
11111111 == -1
00000000 == 0
..

So, if you just subtract 256 == 2**8 from any number that has the most significant bit set, you are done. This is a one-liner.

Code example:

binary = '10000011'
int(binary, 2) - int(binary[0]) * 256

Of course, this can be generalized to n bits:

binary = '101...01'  #<< some n bits. With n==64, if you like.
int(binary, 2) - int(binary[0]) * 2**n

Collectives™ on Stack Overflow

Converting binary representation to signed 64 bit integer in Python

3 Answers 3

Comments

3 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related