How do I get a decimal value when using the division operator in Python?

Question

For example, the standard division symbol '/' rounds to zero:

>>> 4 / 100
0

However, I want it to return 0.04. What do I use?

Solution 1

There are three options:

>>> 4 / float(100)
0.04
>>> 4 / 100.0
0.04

which is the same behavior as the C, C++, Java etc, or

>>> from __future__ import division
>>> 4 / 100
0.04

You can also activate this behavior by passing the argument -Qnew to the Python interpreter:

$ python -Qnew
>>> 4 / 100
0.04

The second option will be the default in Python 3.0. If you want to have the old integer division, you have to use the // operator.

Edit: added section about -Qnew, thanks to ΤΖΩΤΖΙΟΥ!

Solution 2

Other answers suggest how to get a floating-point value. While this wlil be close to what you want, it won't be exact:

>>> 0.4/100.
0.0040000000000000001

If you actually want a decimal value, do this:

>>> import decimal
>>> decimal.Decimal('4') / decimal.Decimal('100')
Decimal("0.04")

That will give you an object that properly knows that 4 / 100 in base 10 is "0.04". Floating-point numbers are actually in base 2, i.e. binary, not decimal.

Solution 3

Make one or both of the terms a floating point number, like so:

4.0/100.0

Alternatively, turn on the feature that will be default in Python 3.0, 'true division', that does what you want. At the top of your module or script, do:

from __future__ import division

Solution 4

You might want to look at Python's decimal package, also. This will provide nice decimal results.

>>> decimal.Decimal('4')/100
Decimal("0.04")

Solution 5

You need to tell Python to use floating point values, not integers. You can do that simply by using a decimal point yourself in the inputs:

>>> 4/100.0
0.040000000000000001

View more solutions

Comments

[Torsten Marek]

Please note that this won't be the case anymore in Python 3.0 if you use /.

[Ishbir]

Please also add the availability of python -Q new command-line option to make your answer more complete.

[Lomithrani]

This gives a floating point value, not a decimal value. See Glyph's answer.

[Mechanical snail]

You can also use from __future__ import division in the source code.

[Jai Narayan]

for python3 you don't require that import division line just use print statements for python2 you need to explicitly add import statement

[Murphy]

Please add details why this would solve the original issue, and perhaps what disadvantages this approach has.

[Vadorequest]

And yet people say python is a great programming language for math... I don't know what's the worse here, using a special import which may change the behaviour if removed, or specifying float operation everywhere.

[Soham Grover]

why does 0.4÷100 give 0.0040000000000000001 instead of just 0.004?

[Glyph]

Because the actual representation of the number is in binary, and since it has to use a finite amount of binary decimal places, the repeating pattern that would produce 4/1000 exactly cannot be represented, in the same way that you cannot represent 1/3 in decimal; 0.33333333333333333 is still less than 1/3; you have to keep going out to infinity. I have written some more aobut this here: blog.glyph.im/2019/10/the-numbers-they-lie.html

[Glyph]

Nowadays, even repr will round this off for you, so you can see the full decimal representation by expanding the precision with an f string: f"{0.004:.100}" results in '0.004000000000000000083266726846886740531772375106811523437‌​5' .