Convert float to int and raise an exception in case of rounding

Question

So for example I have:

a = np.array(...)
a[idx / 3] = 5

idx / 3 will be a float, even though it will always have an integer value in that program. This will raise either a VisibleDeprecationWarning or is already no longer supported by newer Python versions.

Is there a built-in function that converts a float to an int and can raise an Exception if the input value is not an integer? Or what is the proper way to do it?

I'm aware that this might also lead to some questions about 'how close' to an integer, that's why I ask and do not just trust in floor/ceil.

Why is `idx` a float? Looks a bit like a design problem in the first place? — Willem Van Onsem, May 22 '17 at 15:08
@WillemVanOnsem: Because division of an integer by another integer seems to result in a float, even if the number actually has integer value. — Michael, May 22 '17 at 15:22
yes, if you do this with `/` in Python-3. If you use `//` however, it is integer division. — Willem Van Onsem, May 22 '17 at 15:22

score 5 · Accepted Answer · edited May 22 '17 at 15:09

5

Be explicit:

assert (idx % 3) == 0, "idx must be divisible by 3"
a[idx // 3] = 5

edited May 22 '17 at 15:09

Willem Van Onsem

443,496
30
428
555

answered May 22 '17 at 15:08

John Zwinck

239,568
38
324
436

This gives ``IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices`` on NumPy 1.12. You need to actually cast the result of `idx // 3` to an int – Eric May 22 '17 at 16:16
Ah, only if `idx` is a float – Eric May 22 '17 at 16:17
Be carefull, `"assert" statements are removed when the compilation is optimized`, see https://stackoverflow.com/questions/944592/best-practice-for-using-assert/1838411#1838411 – BorjaEst May 28 '23 at 16:59

Convert float to int and raise an exception in case of rounding

1 Answers1