python - Javascript - strip out occurrences of u' in JSON string, parse is returning unexpected token - Stack Overflow

admin管理员组

文章数量:1352810

I have a textfield in a database that contains the results of a python json.dumps(list_instance) operation. As such, the internal fields have a u' prefix, and break the browser's JSON.parse() function.

An example of the JSON string is

"density": "{u'Penobscot': 40.75222856500098, u'Sagadahoc': 
  122.27083333333333, u'Lincoln': 67.97977755308392, u'Kennebec': 
  123.12237174095878, u'Waldo': 48.02117802779616, u'Cumberland':  
  288.9285325791363, u'Piscataquis': 3.9373586457405247, u'Hancock': 
  30.698239582715903, u'Washington': 12.368718341168325, u'Aroostook': 
  10.827378163074039, u'York': 183.47612497543722, u'Franklin':  
  16.89330963710371, u'Oxford': 25.171240748402518, u'Somerset': 
  12.425648288323485, u'Knox': 108.48302300109529, u'Androscoggin': 
  208.75502815768303}"

What I'd like to do is replace those occurrences of u' with a '(single-quote). I've tried

function renderValues(data){
   var pop = JSON.parse(data.density.replace(/u'/g, "'"));
}

but I'm always getting a unexpected token ' exception. Since many of the possible key fields may contain a u, it is not feasable to just delete that character. How can I find all instances of u' and replace with ' without getting the exception?

I have a textfield in a database that contains the results of a python json.dumps(list_instance) operation. As such, the internal fields have a u' prefix, and break the browser's JSON.parse() function.

An example of the JSON string is

"density": "{u'Penobscot': 40.75222856500098, u'Sagadahoc': 
  122.27083333333333, u'Lincoln': 67.97977755308392, u'Kennebec': 
  123.12237174095878, u'Waldo': 48.02117802779616, u'Cumberland':  
  288.9285325791363, u'Piscataquis': 3.9373586457405247, u'Hancock': 
  30.698239582715903, u'Washington': 12.368718341168325, u'Aroostook': 
  10.827378163074039, u'York': 183.47612497543722, u'Franklin':  
  16.89330963710371, u'Oxford': 25.171240748402518, u'Somerset': 
  12.425648288323485, u'Knox': 108.48302300109529, u'Androscoggin': 
  208.75502815768303}"

What I'd like to do is replace those occurrences of u' with a '(single-quote). I've tried

function renderValues(data){
   var pop = JSON.parse(data.density.replace(/u'/g, "'"));
}

but I'm always getting a unexpected token ' exception. Since many of the possible key fields may contain a u, it is not feasable to just delete that character. How can I find all instances of u' and replace with ' without getting the exception?

• javascript
• python
• regex
• json

Share Improve this question Follow asked Jan 23, 2014 at 20:46 JasonJason 11.4k2222 gold badges9393 silver badges185185 bronze badges 6

• Aren't you missing outermost { and } from JOSN string? – anubhava Commented Jan 23, 2014 at 20:56
• @anubhava, no. That is a sample from a serialized return from a Django model. Rather than do an entire data dump, I just showed a sample of the overall data format. – Jason Commented Jan 23, 2014 at 20:58
• Well in that case there is no error when I tried above JSON string with your code. – anubhava Commented Jan 23, 2014 at 20:59
• @anubhava, what browser did you check it with? I'm using chrome 32 on ubuntu. It turns out that replacing the single quotes with double quotes solved the problem. – Jason Commented Jan 23, 2014 at 21:04
• I tested on chrome, can produce jsfiddle also. – anubhava Commented Jan 23, 2014 at 21:08

 |  Show 1 more ment

4 Answers 4

Sorted by: Reset to default 5

The accepted solution is wrong. Your code fails because that string is not valid JSON. Fixing the pseudo-JSON string by replacing it is wrong.

What you have to do is fix the python code that is producing that broken JSON string, which I am pretty sure it is a str() or unicode() where there should be nothing. What you have as a value for the key "density" is a string instead of a dictionary, and therefore, python returns you something like the following:

{"density": "a string that looks like JSON but it is in fact a string reprensentation of a dictionary"}

The function json.dumps will return you always valid JSON strings.

Fix that and you will not have to hack around with filthy string replacements or whatever.

EDIT

Check the following snippet out. There you can see that the u'...' is just the python readable-representation of a unicode object, and has nothing to do whatsoever with a JSON serialization.

>>> import json
>>> data = {u'name': u'Manuel', u'age': 26}
>>> print data
{u'age': 26, u'name': u'Manuel'}  # this is the python representation of a dictionary
>>> print json.dumps(data)
{"age": 26, "name": "Manuel"} # this is a valid JSON string

That JSON is not properly formed, as easy as that.

Updated solution: replace(/u'/g, "'")); => replace(/u'(?=[^:]+')/g, "'"));.

Tested with the following:

"{u'Penobscot': 40.75222856500098, u'Sagadahoc': 122.27083333333333, u'Lincoln': 67.97977755308392, u'Kennebec': 123.12237174095878, u'Waldo': 48.02117802779616, u'Cumberland': 288.9285325791363, u'Piscataquis': 3.9373586457405247, u'Hancock': 30.698239582715903, u'Timbuktu': 12.368718341168325, u'Aroostook': 10.827378163074039, u'York': 183.47612497543722, u'Franklin': 16.89330963710371, u'Oxford': 25.171240748402518, u'Somerset': 12.425648288323485, u'Knox': 108.48302300109529, u'Androscoggin': 208.75502815768303}".replace(/u'(?=[^:]+')/g, "'");

results in:

"{'Penobscot': 40.75222856500098, 'Sagadahoc': 122.27083333333333, 'Lincoln': 67.97977755308392, 'Kennebec': 123.12237174095878, 'Waldo': 48.02117802779616, 'Cumberland': 288.9285325791363, 'Piscataquis': 3.9373586457405247, 'Hancock': 30.698239582715903, 'Timbuktu': 12.368718341168325, 'Aroostook': 10.827378163074039, 'York': 183.47612497543722, 'Franklin': 16.89330963710371, 'Oxford': 25.171240748402518, 'Somerset': 12.425648288323485, 'Knox': 108.48302300109529, 'Androscoggin': 208.75502815768303}"

a little bit old in the answer but if there is no way to change or access the server response try with:

var strExample = {'att1':u'something with u'};

strExample.replace(/u'[\}|\,]/g, "ç'").replace(/u'/g, "'").replace(/ç'/g, "u'");

//{'att1':'something with u'};

The first replace will handle the u' that are in the trailing part of the string in the object changing it to 'ç' character

then removing the u from the phyton unicode and finally change it to u' like the original

I had a similar issue and made this regex that found all of the u's even if the values had them too.

replace(/(?!\s|:)((u)(?='))/g, "")

The accepted answer, I found, missed these occurrences.

I know the OP's doesn't have 'u' for the values and only for keys but thought this may be useful too :)

本文标签：