views:

151

answers:

2

According to the docs, the builtin string encoding string_escape:

Produce[s] a string that is suitable as string literal in Python source code

...while the unicode_escape:

Produce[s] a string that is suitable as Unicode literal in Python source code

So, they should have roughly the same behaviour. BUT, they appear to treat single quotes differently:

>>> print """before '" \0 after""".encode('string-escape')
before \'" \x00 after
>>> print """before '" \0 after""".encode('unicode-escape')
before '" \x00 after

The string_escape escapes the single quote while the Unicode one does not. Is it safe to assume that I can simply:

>>> escaped = my_string.encode('unicode-escape').replace("'", "\\'")

...and get the expected behaviour?

Edit: Just to be super clear, the expected behavior is getting something suitable as a literal.

+1  A: 

Within the range 0 ≤ c < 128, yes the ' is the only difference for CPython 2.6.

>>> set(unichr(c).encode('unicode_escape') for c in range(128)) - set(chr(c).encode('string_escape') for c in range(128))
set(["'"])

Outside of this range the two types are not exchangeable.

>>> '\x80'.encode('string_escape')
'\\x80'
>>> '\x80'.encode('unicode_escape')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
UnicodeDecodeError: 'ascii' codec can’t decode byte 0x80 in position 0: ordinal not in range(128)

>>> u'1'.encode('unicode_escape')
'1'
>>> u'1'.encode('string_escape')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: escape_encode() argument 1 must be str, not unicode

On Python 3.x, the string_escape encoding no longer exists, since str can only store Unicode.

KennyTM
That is just because '\x80' is not a valid ascii encoded string. Try `u'\x80'.encode('unicode-escape')` and you get `'\\x80'`
Mike Boers
@Mike: But is your `my_string` a `str` or a `unicode`?
KennyTM
@KennyTM: unicode
Mike Boers
A: 

According to my interpretation of the implementation of unicode-escape and the unicode repr in the CPython 2.6.5 source, yes; the only difference between repr(unicode_string) and unicode_string.encode('unicode-escape') is the inclusion of wrapping quotes and escaping whichever quote was used.

They are both driven by the same function, unicodeescape_string. This function takes a parameter whose sole function is to toggle the addition of the wrapping quotes and escaping of that quote.

Mike Boers