Search code examples
pythonpython-2.7sign

Why am I getting a positive result with fractional powers?

The following output from IDLE makes no sense to me.

>>> a=-1.0
>>> a**(1/3)
1.0
>>> a==-1.0
True
>>> -1.0**(1/3)
-1.0

Why are two theoretically equivalent statements returning different results? How is Python (2.7) handling the __pow__ method for doubles that this is the result? I just tried it with integers as well and received the same result. Other than computing the sign of the input to the __pow__ function and copying it to the result, how can I fix this?

Solution

  • This is an operator precendence issue:

    >>> -1.0**(1/3)
-1.0
>>> (-1.0)**(1/3)
1.0

    Also, note that (1/3) is zero unless you import division from the __future__, which gives Python 3.x behavior (and a ValueError). Use 1/3. to get 1/3 as a float.