10

I come across a strange problem dealing with python isdigit function.

For example:

>>> a = u'\u2466'
>>> a.isdigit()
Out[1]: True
>>> a.isnumeric()
Out[2]: True

Why this character is a digit?

Any way to make this return False instead, thanks?

Edit, If I don't want to treat it as a digit, then how to filter it out?

For example, when I try to convert it to a int:

>>> int(u'\u2466')

Then UnicodeEncodeError happened.

lxyu
  • 2,661
  • 5
  • 23
  • 29
  • ⑦ is a digit, at least Unicode consortium thinks so. If you have your own understanding of "digits", feel free to share. – georg May 15 '12 at 15:34
  • 2
    What values do you want to treat as digits? If you don't agree with the way `isdigit` treats unicode and just want to look for arabic numerals, you're best off doing `if a in "0123456789"`. – Gort the Robot May 15 '12 at 16:23

3 Answers3

24

U+2466 is the CIRCLED DIGIT SEVEN (⑦), so yes, it's a digit.

If your definition of what is a digit differs from that of the Unicode Consortium, you might have to write your own isdigit() method.

Edit, If I don't want to treat it as a digit, then how to filter it out?

If you are just interested in the ASCII digits 0...9, you could do something like:

In [4]: s = u'abc 12434 \u2466 5 def'

In [5]: u''.join(c for c in s if '0' <= c <= '9')
Out[5]: u'124345'
NPE
  • 486,780
  • 108
  • 951
  • 1,012
9

If you're going to convert something to int you need isdecimal rather than isdigit.

Note that "decimal" is not just 0, 1, 2, ... 9, there are number of characters that can be interpreted as decimal digits and converted to an integer. Example:

#coding=utf8

s = u"1٢٣٤5"
print s.isdecimal() # True
print int(s) # 12345
georg
  • 211,518
  • 52
  • 313
  • 390
4

The character is the CIRCLED DIGIT SEVEN, which is numeric and a digit.

If you want to restrict the digits to the usual 0-9, use a regular expression:

import re

def myIsDigit(s):
  return re.search("[^0-9]", s) is None
Emil Vikström
  • 90,431
  • 16
  • 141
  • 175