将包含转义字符的字符串转换为字典

Question

I need to convert a python string which represents a dict into a python dict. The string might contain any valid dict representation, including windows style pathes (with backslashes), eg

mystring = u'{"baselocaldir":"c:\\tmp\\SrcTmp\\RepManager"}'

I would need a generic str to dict convert function, so this is just an example of a source string, which doesn't work. The source string might come from external sources. A python 2/3 compatible solution would be preferred.

I already tried the given answers:

json.loads does not work (even if I reformat the string to json syntax): raises exception

ast.literal_eval does not work: in this example it places a tab character in the result

eval: the same result as ast.literal_eval

Answer 1

I would put a hack on the string to replace 'c:' as a raw string literal r'c:'

mystring = u'{"baselocaldir": "c:\\tmp\\SrcTmp\\RepManager"}'.replace('"c:', 'r"c:') 
_dict = eval(mystring)
_dict

Result:

{'baselocaldir': 'c:\\tmp\\SrcTmp\\RepManager'}

Answer 2

Edit3: after op change the sample string to double backslash, it is easier and no need to use regex:

mystring = u'{"baselocaldir":"c:\\tmp\\SrcTmp\\RepManager"}'
test = repr(mystring)[1:-1] 
print(test)

# convert to dictionary
my_dict = json.loads(test)
print('dict key "baselocaldir" = ', my_dict["baselocaldir"])

output:

{"baselocaldir":"c:\\tmp\\SrcTmp\\RepManager"}
dict key "baselocaldir" =  c:\tmp\SrcTmp\RepManager

Edit2: apparently using repr() alone is not enough, that's why i edited my answer to use regex and replace all the \\ to \\\\ , here is the code:

import re, json
mystring = u'{"baselocaldir":"c:\tmp\SrcTmp\RepManager"}'

test = re.sub(r'(?<=[^\\])\\(?=[^\\])', r'\\\\', repr(mystring)[1:-1])
print(test)

# convert to dictionary
my_dict = json.loads(test)
print('dict key "baselocaldir" = ', my_dict["baselocaldir"])

output:

{"baselocaldir":"c:\\tmp\\SrcTmp\\RepManager"}
dict key "baselocaldir" =  c:\tmp\SrcTmp\RepManager

previous answer, which is not enough Edit: simple way to convert a string to raw is to use repr() or "%r"

Here is a one step solution, credits goes to Nishanth Amuluru and Jed Alexander 9 years ago:

mystring = u'{"baselocaldir":"c:\tmp\SrcTmp\RepManager"}'

raw_str = "%r"%mystring
rep_str= repr(mystring)

print('original string = ', mystring)
print('Raw string = ', raw_str)
print('rep string = ', rep_str)

output:

original string =  {"baselocaldir":"c:  mp\SrcTmp\RepManager"}
Raw string =  '{"baselocaldir":"c:\tmp\\SrcTmp\\RepManager"}'
rep string =  '{"baselocaldir":"c:\tmp\\SrcTmp\\RepManager"}'

Answer 3

My (maybe not most elegant) solution:

But it works on python2 , python3 and with unicode chars in unicode strings:

text_type = None
if PY2:
    string_types = basestring
    text_type = unicode
else:
    string_types = text_type = str

def DictUnescaceBackslash(oDict):
    for key, value in iteritems(oDict):
        if isinstance(value, dict):
            DictUnescaceBackslash(value)
        elif isinstance(value, string_types):
            oDict[key]=oDict[key].replace("***BaCkSlAsH***","\\")
        elif isinstance(value, list):
           for elem in value:
                DictUnescaceBackslash(elem)

mystring = u'{"baselocaldir":"c:\\tmp\\SrcTmp\\RepManager"}'
uString2 = mystring.replace("\\","***BaCkSlAsH***")
dDict    = ast.literal_eval(uString2)
DictUnescaceBackslash(dDict)