I'm Sure There's A Better Way

[Bengt Richter]

On Sun, 8 Jul 2001 09:40:21 +0200, "Alex Martelli" <aleaxit at yahoo.com>
wrote:

>"Bengt Richter" <bokr at accessone.com> wrote in message
>news:3b478576.1314346961 at wa.news.verio.net...
>    ...
>> >> dollar amount.  Which means it follows these rules:
>> >>
>> >> First character is numeric or "-"
>    ...
>> >This is very easy to express as a Regular Expression:
>> >    r'-?\d+(\.\d\d)?$'
>> or
>>      r'^-?(\d*\.\d\d|\d+)$'
>> if you want to allow .12  (or -.00), see below ;-)
>
>Yes, but that follows a very different spec from the one
>given by the original poster, where the first rule was, as
>above quoted, 'first character is numeric or "-"'.
>
>Getting the specs 'right' for a fuzzily defined domain is
>challenging, I was just trying to help implement the
>specs the poster thought were right:-).
>
>
Well, ok, but did either of us really nail that? ;-)

"""
First character is numeric or "-"
At most, one "." is allowed and, if present, is followed by exactly
two digits
All the remaining characters must be in the range "0" - "9"
"""

I missed because I allowed \.\d\d by itself, which has first
character '.' not equal to '-' or digit.

But according to the rules, ISTM you missed too, since you
don't allow -\.\d\d, which has an ok first character. The rules
don't say whether any digits must precede a '.'

In fact, we probably both missed by not letting '-' by itself
be legal ;-)

Perhaps r'(-|\d)\d*(\.\d\d)?$' ?