Converting JSON-like object to JSON

Question

I have a JSON file that was filtered and lost the original structure. Each line of the file looks like this:

{u'spec1': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property3': u'98754', u'property4': u'val1'}}

json.loads doesn't work and ast.literal_eval doesn't work. I guess I have 2 issues: remove unicode and change ' to ". Can anyone provide some pointers on where to start?

What version of python are you using? I had Python 3.2 and it failed but updating to Python 3.4 works. Python 2.7 also works. — Jeff Mercado
– Jeff Mercado, Commented Jul 29, 2015 at 8:25

liushuaikobe · Accepted Answer · 2015-07-29 08:36:57Z

1

Assuming you are using Python 2.X.

json.loads takes a str or unicode as its param. The string you give is not a valid json string. So we should do some pre-cooking work.

import re, json

json_str = """{u'spec1': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property3': u'98754', u'property4': u'val1'}}"""
json_str = json_str.replace("\'", "\"")
json_str = re.sub(r"u\"", "\"", json_str)

json_dict = json.loads(json_str)

Then the json_dict will be a dictionary inflated from your json string.

edited Jul 29, 2015 at 8:36

answered Jul 29, 2015 at 8:28

liushuaikobe

2,1901 gold badge25 silver badges28 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

user20150316 Over a year ago

{u'spec1': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property3': u'98754', u'property4': u'val1'}} is what I get as a result

liushuaikobe Over a year ago

@user20150316 result of which step?

user20150316 Over a year ago

print json_dict. json_str looks OK. But json_dict has u'

wswld Over a year ago

All this precooking seems a little tedious. What we are having here is a perfectly working Python objects, so why not try to eval them as such, without all that JSON magic?

wswld Over a year ago

Well we could use ast.literal_eval which is a little more stable. Don't get me wrong I think every solution is right as far as it does the job and makes OP happy. I just wanted to note, that there are a bit more elegant ways to do that (an I don't mean mine specifically). Anyways, this discussion has probably run its course. Have a nice day.

|

wswld · Accepted Answer · 2015-07-29 08:50:10Z

Here is my take (Python 2.7).

import StringIO
import ast
file = u"""{u'spec1': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property3': u'98754', u'property4': u'val1'}}
{u'spec2': {u'property1': u'12345', u'property2': 1234}, u'spec3': {u'property3': u'98754', u'property4': u'val1'}}
{u'spec4': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property5': u'98754', u'property4': u'val1'}}
{u'spec6': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property7': u'98754', u'property4': u'val1'}}
"""
buffer = StringIO.StringIO(file)
lines = buffer.readlines()
dicts = []
for line in lines:
    dicts.append(ast.literal_eval(line))
print dicts

Don't look at StringIO, it's there to emulate file-reading. What I'm proposing is to read the file by line and do literal_eval by line.

For me it was the only way to make it work without errors.

airween · Accepted Answer · 2015-07-29 09:44:02Z

1

Looks like this works (in Python3):

$ python3.4
Python 3.4.0 (default, Jun 19 2015, 14:20:21) 
[GCC 4.8.2] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import json
>>> j = """{u'spec1': {u'property1': u'12345', u'property2': 1234}, u'spec2': {u'property3': u'98754', u'property4': u'val1'}}"""
>>> json.loads(j.replace("'","\"").replace('u"','"'))
{'spec1': {'property1': '12345', 'property2': 1234}, 'spec2': {'property4': 'val1', 'property3': '98754'}}

As you can see, I've replaced both ' to " chars, and (thus came) u" to " patterns.

Hope this helps.

a.

edited Jul 29, 2015 at 9:44

answered Jul 29, 2015 at 8:32

airween

6,3231 gold badge16 silver badges23 bronze badges

3 Comments

user20150316 Over a year ago

I am using Python 2.7 and the output still has u'

airween Over a year ago

Of course, this works in Python 2(.7 and below), I just wrote the Python 3, cause' somebody wrote, he uses Python 3, and something is wrong at there. :)

airween Over a year ago

@user20150316: yes, in Python 3, the default encoding of all string is unicode, therefore there isn't any prefix to show, the string is unicode (with "u'"). In Python 2, there isn't, therefore it must be show the prefix, if string is unicode. See: wiki.python.org/moin/Python2orPython3 "The most drastic improvement is the better Unicode support (with all text strings being Unicode by default)..."

Collectives™ on Stack Overflow

Converting JSON-like object to JSON

3 Answers 3

7 Comments

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

7 Comments

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related