method to remove zero digits from fraction in decimal

[bowlofred]

i want this to be in the decimal.Decimal type which may or may not store the fraction, which can be zero if stored.

sorry, my first post just said "decimal" instead of "decimal.Decimal".

I don't know anything in Decimal that will do this. But the same conversion algorithm would still work. If it is equivalent to the integer, reset it.

>>> l = [decimal.Decimal(f"{d}.0") / decimal.Decimal("2") for d in range(1,11)]
>>> [str(x) for x in l]
['0.5', '1.0', '1.5', '2.0', '2.5', '3.0', '3.5', '4.0', '4.5', '5.0']
>>> l = [decimal.Decimal(int(d)) if d == int(d) else d for d in l]
>>> [str(x) for x in l]
['0.5', '1', '1.5', '2', '2.5', '3', '3.5', '4', '4.5', '5']

[Skaperen]

i have coded:

...
if number == int(number)
    number = int(number)
...

just like you suggested above in post #2. i was looking for something i could put in the print() call and eliminate those 2 lines. i always like shorter code as it make it easier to read (and faster to compile on Thursdays).

if i had been the one who implemented decimal, i would have included such a method. i don't know what i would have named it.

edit:

such a method would return the modified value, as Decimal objects are immutable.

[Skaperen]

i coded this little thing:

from decimal import Decimal
point = str(3/2)[1]

def dec_min_frac(n):
    if isinstance(n,(int,float,str,bytes,bytearray)):
        n = Decimal(n)
    elif isinstance(n,list):
        return [dec_min_frac(x)for x in n]
    elif isinstance(n,tuple):
        return tuple(dec_min_frac(x)for x in n)
    else:
        return None
    s = str(n)
    if point in s:
        while s[-1] == '0':
             s = s[:-1]
        if s[-1] == '.':
             s = s[:-1]
    return Decimal(s)


if __name__ == '__main__':
    from sys import argv
    argv.pop(0)
    print(' '.join(str(x) for x in dec_min_frac(argv)))

[wavic]

In [1]: import math

In [2]: x = lambda num: int(f[1]) if (f := math.modf(num))[0] == 0 else f[0] + f[1]

In [3]: x(4.0)
Out[3]: 4

In [4]: x(4.1)
Out[4]: 4.1

Just noticed you are talking about decimal.Decimal numbers.
It works with those too. Just tested

In [5]: import decimal

In [6]: num = decimal.Decimal('37')

In [7]: x(4.1)
Out[7]: 4.1

In [8]: x(4.0)
Out[8]: 4

In [9]: x(num)
Out[9]: 37

In [10]: num = decimal.Decimal('37.0')

In [11]: x(num)
Out[11]: 37

In [12]: num = decimal.Decimal('37.1')

In [13]: x(num)
Out[13]: 37.1

Ignore in/out 7 and 8. I repeated them by mistake

[wavic]

Well, I just looked at decimal module docs.

I think the normalize method is what you need.

In [1]: import decimal

In [2]: num = decimal.Decimal('37.0')

In [3]: num.normalize()
Out[3]: Decimal('37')

In [4]: num = decimal.Decimal('37.1')

In [5]: num.normalize()
Out[5]: Decimal('37.1')

In [6]: num = decimal.Decimal('37')

In [7]: num.normalize()
Out[7]: Decimal('37')

[Skaperen]

I think the normalize method is what you need.

the library manual (PDF version for 3.7) describes conversion to what appears to be an exponential form, e.g. Decimal('0e0'). that's not what i wanted so i skipped further consideration of that method.

edit:

testing shows that the manual seems to be wrong, incomplete, or misleading. it looks like .normalize() may well be useful.

[wavic]

One line after their Decimal('0') to Decimal('0e0') explanation there is an example that says

Decimal('32.100') and Decimal('0.321000e+2') both normalize to the equivalent value Decimal('32.1').

Both manuals ( 3.7 and 3.10 ) are the same. Perhaps you have been in a hurry.

Morning is cold

[Skaperen]

Both manuals ( 3.7 and 3.10 ) are the same. Perhaps you have been in a hurry.

i have not downloaded any manual since 3.7, which i did when i was running 3.6, taking care to check for changes made as done in 3.7). thus, i have no 3.10 handy to compare with. i use the PDF versions.

maybe i should go download manuals for 3.8, which is the system Python version in {U,Xu}buntu 20.04 LTS which makes use of Python so i am reluctant to consider upgrading to a newer version unless it comes from Canonical.