why does math.pow yields OverflowError (while python itself cancalculate that large number)

[Tzury Bar Yochay]

What is the reason math.pow yields OverflowError while python itself
can
calculate these large numbers. e.g:

>>import math
math.pow(100, 154)

1e+308

>>math.pow(100, 155)

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
OverflowError: math range error

>>eval(('100*'* 155)[:-1])

10000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
0000000000000000L

>>>

[John Machin]

On Oct 23, 8:21*pm, Tzury Bar Yochay <[e-mail]:

What is the reason math.pow yields OverflowError while python itself
can
calculate these large numbers. e.g:

>import math
math.pow(100, 154)

1e+308

>math.pow(100, 155)

Traceback (most recent call last):
* File "<stdin>", line 1, in <module>
OverflowError: math range error>>eval(('100*'* 155)[:-1])

10000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
0000000000000000L

Because math.pow returns a float; 100 ** 155 won't fit in a float.

[Tzury Bar Yochay]

Because math.pow returns a float; 100 ** 155 won't fit in a float.

Sure that is the reason.
May I rephrase, my question:
Why not returning another type as long as we can calculate it?
After all, math module is likely to be used on large numbers as well.

[Christian Heimes]

Tzury Bar Yochay wrote:

>Because math.pow returns a float; 100 ** 155 won't fit in a float.

Sure that is the reason.
May I rephrase, my question:
Why not returning another type as long as we can calculate it?
After all, math module is likely to be used on large numbers as well.

Because it's very complicated to get it right. The math module is a thin
wrapper around the platform math library. It took Mark and me very long
to get cmath and math right for floats (C doubles). Arbitrary precision
and arbitrary length numbers are a different story. You have to use 3rd
party tools for those. The standard library aims to solve common
problems, not special problems.

Christian

[John Machin]

On Oct 23, 9:24*pm, Tzury Bar Yochay <[e-mail]:

Because math.pow returns a float; 100 ** 155 won't fit in a float.

Sure that is the reason.
May I rephrase, my question:
Why not returning another type as long as we can calculate it?
After all, math module is likely to be used on large numbers as well.

The math module is intended to replicate the functionality found in
math.h in the C Standard Library; that's it, no more, no less. There
are other libraries if you want more-than-float precision.

[Marc 'BlackJack' Rintsch]

On Thu, 23 Oct 2008 02:21:36 -0700, Tzury Bar Yochay wrote:

>>>eval(('100*'* 155)[:-1])

10000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
00000000000000000000000000000000000000000000000000 000000000
0000000000000000L

>>>>

This can be written more straigth forward as ``100**155`` or
``pow(100, 155)``. No need for `eval()`\ing a string.

Ciao,
Marc 'BlackJack' Rintsch