Make dictionary from dictionary string

Question

I made a dictionary from one file and need to write it into another file that I will be making edits to. To do this, I made a dictionary from the first file and made it into a string to write it to a second file. Is there a way to convert this string back into a dictionary?

An example of the first file is:

123 2
125 5
128 3

which I make into a dictionary and make into a string in the second file:

def CreateNew(prefix, old_file):
    new_file = open('new_file_name', 'w')
    new_file.write(str(old_file))
    new_file.close()
    return new_file

Now, I need to make edits to some of the values in this new file, but am not sure how to make this dictionary string back into a dictionary. I wanted to just print this off to test it out:

def EmpTrans(transactions_file, new_file):
    print(dict(new_file))

but this gives me an empty dictionary {}.
I'm trying not to use any modules for this. I was able to use eval().

Answer 1

To print a dictionary to a file:

output_file = open(path_to_stored_dictionary, 'w')
output_file.write(str(my_dictionary))
output_file.close()

To read a dictionary from a file:

my_dictionary = open(path_to_stored_dictionary, 'r').read()
my_dictionary = eval(my_dictionary)

Note @TigerhawkT3's comment:

...eval() executes code, which can be dangerous if untrusted sources will be sending strings to this function.

Answer 2

To turn your string back into a dictionary, you would split the string into a tuple of key, value pairs. To do so, you would split it.

def get_dict(s):
    return dict(map(lambda kv: kv.split(" "), s.split("\n")))

But I would advise against this. In this case, you should use the Pickle module, except if you don't absolutely trust the data. See also: How can I use pickle to save a dict?.

import pickle

def dump(dict_, file):
    with open(file + ".pickle", "wb") as handle:
        pickle.dump(dict_, handle)

def load(file):
    with open(file + ".pickle", "rb") as handle:
        return pickle.load(handle)

The dict_ can really be any object.

Answer 3

Another secure function is using JSON. Here i use the json.dump() and json.load() functions. The only handicap, is that json load returns strings as unicode string, therefor i am using an own JSONDecoder class that calls super and then traverses the result to encode the strings:

import json
from types import *

# your own decoder class to convert unicode to string type
class MyJSONDecoder(json.JSONDecoder):

    # decode using our recursive function
    def decode(self,json_string):
        return self._decode_obj(super(MyJSONDecoder,self).decode(json_string))

    # recursive function to traverse lists
    def _decode_list(self,data): 
        decoded = []
        for item in data:
            if type(item) is UnicodeType:   item = item.encode('utf-8')
            elif type(item) is ListType:    item = self._decode_list(item)
            elif type(item) is DictType:    item = self._decode_obj(item)
            decoded.append(item)
        return decoded

    # recursive function to traverse objects
    def _decode_obj(self,data): 
        decoded = {}
        for key, value in data.iteritems():
            if type(key) is UnicodeType:    key = key.encode('utf-8')
            if type(value) is UnicodeType:  value = value.encode('utf-8')
            elif type(value) is ListType:   value = self._decode_list(value)
            elif type(value) is DictType:   value = self._decode_obj(value)
            decoded[key] = value
        return decoded

# the dictionary to save
dict = {
    "123": 2,
    "125": 4,
    "126": 5,
    "128": 6
}

# decoder instance
my_decoder = MyJSONDecoder()

# write object to file
with open('serialized.txt', 'w') as new_file:
    json.dump(dict, new_file)
    new_file.close()
    print "Original", dict

# read object from file
with open ("serialized.txt", "r") as old_file:
    dictcopy = my_decoder.decode(old_file.read())
    old_file.close()
    print "**Copy**", dictcopy