Python binascii.Error: Incorrect padding [Solved]

# Table of Contents

# Python binascii.Error: Incorrect padding [Solved]

The Python "binascii.Error: Incorrect padding" occurs when you try to base64 decode a bytes object that doesn't have the correct padding.

To solve the error add padding to the bytes object when calling base64.b64decode().

binascii error incorrect padding in python

Here is an example of how the error occurs.

main.py

Copied!
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='


decoded_bytes = base64.b64decode(encoded_bytes)
print(decoded_bytes)  # 👉️ b'{"name": "bobby hadz"}'

# ⛔️ binascii.Error: Incorrect padding
print(base64.b64decode(encoded_bytes[:-2]))

We used the base64.b64encode() method to encode a bytes-like object using Base64.

The method returns the encoded bytes.

We then used the base64.b64decode() method to decode the Base64 encoded bytes-like object.

The last line raises the "binascii.Error: Incorrect padding" error because we used slicing to remove the last 2 characters ( the padding - ==) from the bytes object.

# Add padding to your bytes object before decoding

You can solve the error by adding padding (==) to your bytes object before decoding.

The base64.b64decode() method will truncate any extra padding, so even if your bytes object is padded sufficiently, it won't cause any issues.

main.py

Copied!
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='


decoded_bytes = base64.b64decode(encoded_bytes)
print(decoded_bytes)  # 👉️ b'{"name": "bobby hadz"}'

# 👇️ b'{"name": "bobby hadz"}'
print(base64.b64decode(encoded_bytes[:-2] + b'=='))

# 👇️ True
print(data == base64.b64decode(encoded_bytes[:-2] + b'=='))

The code for this article is available on GitHub

You can add padding to your bytes object by using the addition (+) operator.

main.py

Copied!
# 👇️ b'{"name": "bobby hadz"}'
print(base64.b64decode(encoded_bytes[:-2] + b'=='))

pad bytes object before base64 decoding

The base64.b64decode() method truncates the extra padding, so the following code sample achieves the same result.

main.py

Copied!
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='


decoded_bytes = base64.b64decode(encoded_bytes)
print(decoded_bytes)  # 👉️ b'{"name": "bobby hadz"}'

# 👇️ b'{"name": "bobby hadz"}'
print(base64.b64decode(encoded_bytes[:-2] + b'======='))

# 👇️ True
print(data == base64.b64decode(encoded_bytes[:-2] + b'========='))

pad bytes object extra

You can simply add the maximum number of pad characters that you need.

main.py

Copied!
# 👇️ b'{"name": "bobby hadz"}'
print(base64.b64decode(encoded_bytes[:-2] + b'======='))

If you don't want to add extra padding, use the following formula.

main.py

Copied!
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='


bytes_without_padding = encoded_bytes[:-2]

# 👇️ b'{"name": "bobby hadz"}'
print(base64.b64decode(bytes_without_padding +
      b'=' * (-len(bytes_without_padding) % 4)))

# 👇️ True
print(data == base64.b64decode(bytes_without_padding +
      b'=' * (-len(bytes_without_padding) % 4)))

The code for this article is available on GitHub

However, using the formula is not necessary and is a bit more difficult to read.

using formula to add padding to bytes

# Make sure the `validate` keyword argument is set to `False`

Make sure the validate keyword argument is set to False when calling the base64.b64decode() method.

main.py

Copied!
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='


decoded_bytes = base64.b64decode(encoded_bytes)
print(decoded_bytes)  # 👉️ b'{"name": "bobby hadz"}'

# 👇️ b'{"name": "bobby hadz"}'
print(
    base64.b64decode(
        encoded_bytes[:-2] + b'=======',
        validate=False
    )
)

# 👇️ True
print(data == base64.b64decode(
    encoded_bytes[:-2] + b'=========',
    validate=False
)
)

The code for this article is available on GitHub

If your bytes-like object already has some padding, you have to set the validate keyword argument to False.

If validate is set to True, a binascii.Error error is raised if the padding is longer than 2 characters.

When the validate argument is set to False, the b64decode method truncates any extra padding automatically.

The default value for the validate keyword argument is False, so not passing it and setting it to False explicitly is equivalent.

# Creating a reusable function that pads the bytes

You can also create a reusable function that pads the bytes and base64 decodes them.

main.py

Copied!
import re
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='


def decode_base64(encoded_bytes, altchars=b'+/'):
    encoded_bytes = re.sub(
        rb'[^a-zA-Z0-9%s]+' %
        altchars, b'', encoded_bytes)

    missing_padding_length = len(encoded_bytes) % 4

    if missing_padding_length:
        encoded_bytes += b'=' * (4 - missing_padding_length)

    return base64.b64decode(encoded_bytes, altchars)


bytes_without_padding = encoded_bytes[:-2]

# 👇️ b'{"name": "bobby hadz"}'
print(decode_base64(bytes_without_padding))

The code for this article is available on GitHub

The decode_base64 function:

Takes an encoded bytes object as a parameter.
Pads the bytes object sufficiently, so there is no missing padding.
Base64 decodes the bytes object.

The re.sub() method is used to normalize the bytes object.

The method removes any non-Base64 characters and returns the result.

# Removing the prefix when decoding base64 image strings

If your string starts with a prefix of data:image/png;base64,, you have to remove it before base64 decoding it

main.py

Copied!
from io import BytesIO
import base64
from PIL import Image

data = 'data:image/png;base64, iVBORw0KGgoAAAANSUhEUgAAAAUAAAAFCAYAAACNbyblAAAAHElEQVQI12P4//8/w38GIAXDIBKE0DHxgljNBAAO9TXL0Y4OHwAAAABJRU5ErkJggg=='


im = Image.open(BytesIO(base64.b64decode(data.split(',')[1])))
im.save("my-image.png")

The code for this article is available on GitHub

We used the str.split method to split the string on the comma and then accessed the element at index 1 (the second list element).

This effectively removes the data:image/png:base64 prefix from the string.

We can then pass the result to the base64.b64decode() method.

Make sure you have the Pillow module installed to be able to run the code sample.

shell

Copied!
pip install Pillow

# 👇️ or with pip3
pip3 install Pillow

# Try using the `base64.urlsafe_b64decode` method instead

If the error persists, try to use the base64.urlsafe_b64decode method instead.

main.py

Copied!
import base64

data = bytes('{"name": "bobby hadz"}', encoding='utf-8')

encoded_bytes = base64.b64encode(data)
print(encoded_bytes)  # b'eyJuYW1lIjogImJvYmJ5IGhhZHoifQ=='

decoded_bytes = base64.urlsafe_b64decode(encoded_bytes)
print(decoded_bytes) # 👉️ b'{"name": "bobby hadz"}'

The code for this article is available on GitHub

The base64.urlsafe_b64decode() method decodes the supplied bytes-like object or ASCII string using the URL and filesystem-safe alphabet.

The method substitutes + with - and / with _ in the standard Base64 alphabet.

The method returns the decoded bytes.

# Additional Resources

You can learn more about the related topics by checking out the following tutorials:

I wrote a book in which I share everything I know about how to become a better, more efficient programmer.

You can use the search field on my Home Page to filter through all of my articles.

Python binascii.Error: Incorrect padding [Solved]

# Table of Contents