Convert a string into a raw string (Python recipe) by Brett Cannon

ActiveState Code (http://code.activestate.com/recipes/65211/)

\0 through \9 are not great examples. The example dictionary should not contain codes like r'\0'. While the letter escapes work, you need to represent the numeric codes as full 3-digit octal (r'\000') or hex (r'\x00') values, or you will be surprised translating the two-character string "\x007' into a constant that becomes a bell character. '\8' and '\9' are really just the characters '8' and '9'. It is also a bad idea to have a dictionary constant with conflicting entries ('\7' and '\a').

So, replace the last part of the escape_dict definition with:

...,
'\"':r'\"',
'\0':r'\000',
'\1':r'\001',
'\2':r'\002',
'\3':r'\003',
'\4':r'\004',
'\5':r'\005',
'\6':r'\006'}

Defeats Purpose. If I use your suggestion, I end up with the following result:

>>> raw('\1\2\3')
'\\001\\002\\003'
>>> r'\1\2\3'
'\\1\\2\\3'

The whole point of this is to get back the string to the form it should be in had an r been appended to the string when it was created. Unfortunately your solution breaks that.

To convert a normal string that is generated during runtime into a raw string...

>>> plain_str = 'newline \n'
>>> raw_str = r'newline\n'
>>> plain_str == raw_str
False
>>> temp_str = "%r"%plain_str
>>> print temp_str
'newline \\n'
>>> new_raw_str = temp_str[1:-1]
>>> new_raw_str == raw_str
True

@Nishanth Thank you. I scoured the web for an hour looking for this answer. I knew that I had to convert the runtime string to a literal, just was unsure about how to do it.

After doing some research to determine what the "%r" was doing in the format string above, this turns out to be really easy. The built-in function repr() does exactly this.

@Nishanth: that doesn't convert a plain string to a raw string--it converts to the "printable representation" of the plain string, and then chops off the quotes.

see this counter-example:

>>> plain = "\\a\a\a\b"
>>> raw = r"\\a\a\a\b"
>>> print plain
\a
>>> print raw
\\a\a\a\b
>>> plain == raw
False
>>> temp = "%r" % plain
>>> new = temp[1:-1]
>>> print new
\\a\x07\x07\x08
>>> print raw
\\a\a\a\b
>>> new == raw
False
>>>

I'm trying to figure out how to "rawify" strings containing backslashes that I'm reading from stdin and need to insert into a database. I'll continue looking and try to report back with what I find.

any further help would be wonderful!

I use the following method to convert a python string (str or unicode) into a raw string:

def raw_string(s):
    if isinstance(s, str):
        s = s.encode('string-escape')
    elif isinstance(s, unicode):
        s = s.encode('unicode-escape')
    return s

Example usage:

import re

s = "This \\"
re.sub("this", raw_string(s), "this is a text")

This works great :)

@Ratnadeep Debnath Not exactly. If I want to use '\1foo\2' it gets translated to '\x01foo\x02' which is not the same as r'\1foo\2', and doesn't work with re.sub() properly. Any solution to that?