Character to decimal conversion

This converts from character representation of integers to their decimal value:

def chrtodec(str):
    dec = 0
    base = ord('0')
    for chr in str:
        dec = dec * 10 + ord(chr) - base
    return dec

For example: "123" -> 123

You shouldn't use chr or str as names because they shadow the builtin chr and str methods. If you then wanted to use str() you'd be out of luck because str is now a string instead of a function. Given that your function deals with both strings and character ordinals it is not unlikely that these could be used. string is a potential improvement, it can occasionally cause trouble if you're trying to use the string module of the same name.

You should also add comments and a docstring. Docstrings are basically comments that are programmatically accessible so that other users can understand how to use your function.

def chrtodec(str):
    """Converts a string to a float using character ordinal positions."""

I think you have another big problem though, which is invalid input handling. Take a look at these:

>>> chrtodec("hello")
619663
>>> chrtodec("12.21")
11821
>>> chrtodec("523.32f")
5228374

Clearly there are problems here when you have characters other than numbers, so I think what you should do is raise a ValueError when an invalid string is passed. You already have the tools to do this figured out of course if you just check that a character's ordinal fits in the right range.

if not (ord('0') <= ord(chr) <= ord('9')):
    raise ValueError("Invalid character {} in string {},"
                     "only digits are parseable.".format(chr, str))

You could also just use chr.isdigit() as @Barry pointed out.

Also you could add support for negative numbers by checking at the start if the first character is ord('-').

negative = ord(str[0]) == ord('-')

This evaluates the expression str[0] == ord('-') and sets negative as the boolean result. Note that to make this compatible with the error handling I suggested, you should then remove the first character from str. And probably update the error message and docstring too.

if negative:
    str = str[1:]

Then just return with a ternary that checks negative.

return dec if not negative else -dec

StackExchange Code Review Q#104289, answer score: 6