Question

I observed this strange behaviour of center method of str objects in python 2.7 –

>>> print '-'.center(5,'.')        # '-' is the minus or hyphen key on the keyboard.
..-..                              # correctly centered
>>> print '─'.center(5, '.')       # '─' is NOT the minus or hyphen key on the keyboard.
.─.                                # incorrectly centered

Is this a bug in the implementation of center method or has something to do with the use of unicode characters?
How do I fix this?

[Note that this works perfectly fine in python 3.2]

Answer 1

In Python 2, there’s a distinction between the str and unicode types. A str is a sequence of bytes and is created with un-modified quotes. Your '─' is creating a sequence of 3 bytes which display as a single unicode character; when you ask str.center() to center those bytes in a field 5 bytes wide, it adds a single byte on either side; when printed you only get 3 glyphs.

If you create a unicode object (with u'─' or ‘─’.decode(‘utf-8’)), you now have a sequence of 1 unicode character, which will be centered by adding 2 characters to either side.

Python 3 doesn’t have this problem, because ordinary strings are unicode character sequences.