A useful, but painful, one-liner to edit money amounts

[John Nagle]

There's got to be a better way to do this:

def editmoney(n) :
return((",".join(reduce(lambda lst, item : (lst + [item]) if
item else lst,
re.split(r'(\d\d\d)',str(n)[::-1]),[])))[::-1])

>>> editmoney(0)
'0'
>>> editmoney(13535)
'13,535'
>>> editmoney(-14535)
'-14,535'
>>> editmoney(123456)
'123,456'
>>> editmoney(1234567890)
'1,234,567,890'
>>> editmoney(-1234)
'-1,234'

The basic idea here is that we want to split the string of digits
into groups of 3 digits, aligned at the right. Because regular
expressions are right to left, we have to reverse the string to
do that, then reverse again at the end. s[::-1} reverses an
interable.

"split" with a capturing group introduces empty strings into the
list. Hence the "reduce" and lambda to get rid of them.

Any better ideas?

(Yes, I know there's a built-in feature for this scheduled for
Python 2.7.)

John Nagle

[Paul Rubin]

John Nagle <na...@animats.com> writes:
> def editmoney(n) :
> return((",".join(reduce(lambda lst, item : (lst + [item]) if
> item else lst,
> re.split(r'(\d\d\d)',str(n)[::-1]),[])))[::-1])

Too obscure. I usually use something like this:

def editmoney(n):
if n < 0: return '-' + editmoney(-n)
if n >= 1000:
return editmoney(n // 1000) + ',%03d'% (n % 1000)
return '%d'% n

[Steven D'Aprano]

On Wed, 04 Aug 2010 21:33:31 -0700, John Nagle wrote:

> There's got to be a better way to do this:
>
>
> def editmoney(n) :
> return((",".join(reduce(lambda lst, item : (lst + [item]) if
> item else lst,
> re.split(r'(\d\d\d)',str(n)[::-1]),[])))[::-1])

What does the name "editmoney" mean?

Why the obfuscated one-liner? It's not like you're using it in-line,
you're putting it in a function, so who cares if it's one line or twenty?

def group_digits(n, size=3, sep=','):
"""Group int n in groups of size digits separated by sep."""
s = str(n)
m = len(s) % size
head = s[0:m]
tail = s[m:]
groups = [tail[i*size:(i+1)*size] for i in range(len(tail)//size)]
tail = sep.join(groups)
if head and tail:
return head + sep + tail
elif tail:
return tail
else:
return head

>>> group_digits(0)
'0'
>>> group_digits(1234567890)
'1,234,567,890'
>>> group_digits(1234567890, 4, ';')
'12;3456;7890'

Additional error checking, a better docstring, and extending it to
support negative numbers is left as an exercise.

--
Steven

[Peter Otten]

John Nagle wrote:

>>> locale.setlocale(locale.LC_ALL, ("en_US", "UTF-8"))
'en_US.UTF8'
>>> print locale.currency(13535, grouping=True)
$13,535.00
>>> print locale.format("%d", 13535, grouping=True)
13,535

>>> locale.setlocale(locale.LC_ALL, "")
'de_DE.UTF-8'
>>> print locale.currency(13535, grouping=True)
13.535,00 €
>>> print locale.format("%d", 13535, grouping=True)
13.535

Peter

[geremy condra]

I had literally no idea this existed. Thanks.

Geremy Condra

[Steven D'Aprano]

On Thu, 05 Aug 2010 00:22:57 -0700, geremy condra wrote:

>>>>> locale.setlocale(locale.LC_ALL, "")
>> 'de_DE.UTF-8'
>>>>> print locale.currency(13535, grouping=True)
>> 13.535,00 €
>>>>> print locale.format("%d", 13535, grouping=True)
>> 13.535
>>
>> Peter
>
> I had literally no idea this existed. Thanks.

I knew it existed, but completely forgot about it.

Thanks also Peter.

--
Steven

[Chris Withers]

Peter Otten wrote:
>>>> locale.setlocale(locale.LC_ALL, ("en_US", "UTF-8"))
> 'en_US.UTF8'
>>>> print locale.currency(13535, grouping=True)
> $13,535.00

Okay, so if I'm writing a wsgi app, and I want to format depending on
the choices of the currently logged in users, what would you recommend?

I can't do setlocale, since that would affect all users, and in a
mult-threaded environment that would be bad.

Does that mean the whole locale package is useless to all web-app builders?

Chris

[DarkBlue]

from re import *

class editmoney(float):
def __init__(self,mymoney):
self.mymoney = mymoney
def __str__(self):
temp = "%.2f" % self.mymoney
profile = compile(r"(\d)(\d\d\d[.,])")
while 1:
temp, count = subn(profile,r"\1,\2",temp)
if not count: break
return temp

[Peter Otten]

Chris Withers wrote:

Blame it on the C guys ;)

I've seen

http://babel.edgewall.org/wiki/Documentation/intro.html
http://babel.edgewall.org/wiki/ApiDocs/babel.numbers
http://babel.edgewall.org/wiki/BabelFaq#WhatalternativesexistforPythonprojects

mentioned here but not yet tried it myself.

Peter

[John Posner]

On 8/5/2010 12:33 AM, John Nagle wrote:
> There's got to be a better way to do this:
>
>
> def editmoney(n) :
> return((",".join(reduce(lambda lst, item : (lst + [item]) if
> item else lst,
> re.split(r'(\d\d\d)',str(n)[::-1]),[])))[::-1])
>

Here's a more elegant variant, using regexp lookahead:

def thous_format(integer_string):
"""
add comma thousands separator(s) to an integer-valued string
"""
return re.sub(r'(\d{3})(?=\d)', r'\1,', integer_string[::-1])[::-1]

I *thought* that I had found this on python-list on or about July 5, but
I didn't find the thread after a search through the archives.

-John

[MRAB]

You don't need to reverse the string:

def thous_format(integer_string):
"""
add comma thousands separator(s) to an integer-valued string
"""

return re.sub(r"(?<=\d)(?=(?:\d\d\d)+$)", ",", integer_string)

[John Posner]

On 8/5/2010 12:36 PM, MRAB wrote:

> You don't need to reverse the string:
>
> def thous_format(integer_string):
> """
> add comma thousands separator(s) to an integer-valued string
> """
> return re.sub(r"(?<=\d)(?=(?:\d\d\d)+$)", ",", integer_string)

Nice! My first encounter with a look-behind! It appears that a slight
simplification is possible:

return re.sub(r"(?<=\d)(?=(\d\d\d)+$)", ",", integer_string)

There no harm in putting \d\d\d into a group, though the group is never
used.

-John

[John Posner]

On 8/5/2010 12:36 PM, MRAB wrote:

> You don't need to reverse the string:
>
> def thous_format(integer_string):
> """
> add comma thousands separator(s) to an integer-valued string
> """
> return re.sub(r"(?<=\d)(?=(?:\d\d\d)+$)", ",", integer_string)

Nice! My first encounter with a look-behind! It appears that a slight

[Peter Otten]

Steven D'Aprano wrote:

You're welcome!

[Rebel Lion]

In [1]: '{:,}'.format(-12345678)
Out[1]: '-12,345,678'

[Chris Withers]

DarkBlue wrote:
> On Aug 5, 7:06 pm, Chris Withers <ch...@simplistix.co.uk> wrote:
>> Peter Otten wrote:
>>>>>> locale.setlocale(locale.LC_ALL, ("en_US", "UTF-8"))
>>> 'en_US.UTF8'
>>>>>> print locale.currency(13535, grouping=True)
>>> $13,535.00
>> Okay, so if I'm writing a wsgi app, and I want to format depending on
>> the choices of the currently logged in users, what would you recommend?
>>
>> I can't do setlocale, since that would affect all users, and in a
>> mult-threaded environment that would be bad.
>>
>> Does that mean the whole locale package is useless to all web-app builders?
>>
>> Chris
>
> from re import *
>
> class editmoney(float):
> def __init__(self,mymoney):

What does this have to do with my question about locales in
multi-threaded environments?

Chris

--
Simplistix - Content Management, Batch Processing & Python Consulting
- http://www.simplistix.co.uk