11

While working on an issue, I stumbled at something I cannot really undestand on my own.

I have a variable: a = pow(2, 1024)

Its type is long int. If I try casting it explicitly to float, like float(a), I receive OverflowError. The number is too big to fit in 64bit float, so it's understandable.

Then I try implicit cast, through mutliplying it by a float number:

b = a * 11.0

Once again, the OverflowError occurs, which is fine, because according to python docs, implicit conversion from long int to float happens. So the result is like before.

Finally, I try comparison:

a > 11.0 returns True. The OverflowError doesn't occur. And that confuses me a lot. How does the Python comparison mechanism work, if it doesn't require numerics to be in the same format? Accordind to this,

Python fully supports mixed arithmetic: when a binary arithmetic operator has operands of different numeric types, the operand with the “narrower” type is widened to that of the other, where plain integer is narrower than long integer is narrower than floating point is narrower than complex. Comparisons between numbers of mixed type use the same rule. The constructors int(), long(), float(), and complex() can be used to produce numbers of a specific type.

My question is, why a is not being cast to float in forementioned comparison?

The version of Python I'm using is 2.7.15. Thanks in an advance

Improve this question
1

2 Answers 2

Reset to default
9

From the source:

/* Comparison is pretty much a nightmare.  When comparing float to float,
 * we do it as straightforwardly (and long-windedly) as conceivable, so
 * that, e.g., Python x == y delivers the same result as the platform
 * C x == y when x and/or y is a NaN.
 * When mixing float with an integer type, there's no good *uniform* approach.
 * Converting the double to an integer obviously doesn't work, since we
 * may lose info from fractional bits.  Converting the integer to a double
 * also has two failure modes:  (1) a long int may trigger overflow (too
 * large to fit in the dynamic range of a C double); (2) even a C long may have
 * more bits than fit in a C double (e.g., on a 64-bit box long may have
 * 63 bits of precision, but a C double probably has only 53), and then
 * we can falsely claim equality when low-order integer bits are lost by
 * coercion to double.  So this part is painful too.
 */

As such, the potential pitfalls of conversion are taken into account.

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

5 Comments

but this does not say how it is actually done.. Only that it is painful..
How it is done is in the source code just below that comment. Particularly around line 484.
Uh, that really is painful. Would the case be the same for PyPy as well?
Probably, but is it actually an issue? None of these edge cases break the intuitive comparison operators. If you need to check that it fits a float, then do a float(a) and catch the exception.
I needed my code to be as little exception-driven as possible. Anyway, I've changed the way it works so now I don't happen to have a long int that exceeds float. I will mark your post as an answer.
3

The exact error is OverflowError: int too large to convert to float

Which also means that any int that generates that error is by definition bigger then any possible float. So just checking if it's bigger should return True.

I'm not entirely sure, but I wouldn't be surprised if the implementation is just catching this error in the background (when trying to cast to float) and returning True in that case.

That is always true with the exception of float('inf') which is a special case that should return False (and does)

Improve this answer

5 Comments

this is just a guess
@Chris_Rands I'm not sure if that's the exact implementation but that's the logic of it, because the fact is - it works (as OP mentioned)
float(-a) also triggers OverflowError. -a > 11.0 return false. I think that's not the case
But in that case one is clearly positive and one is negative, there's no reason to check anything else besides that (with comparing to 0 of both elements).
Again, i'm sure by description isn't accurate but it's a logical step to understanding what is actually going on

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.