How to strip invalid characters from JSON output in PHP

Question

I am a PHP developer working on code that will step through and parse a JSON object so that I can manipulate data before outputting it in its final form. Here's some example JSON (its very large so I am removing most of the items I need and just putting the invalid part):

[
    {
        "items": [
            {
                // this is valid and what I need.
            }
        ],
        "espots": [
            {
                "content": "
[TopNavigation_Appliances]
"
            }
        ]
    },

I have effectively stripped the JSON from its source to be strictly the JSON data that's being passed to json_decode(). However, I am always getting a null result because the JSON object has an invalid part that my code doesn't actually require to function. The whole "espots" part of the JSON is unneeded within my application, but the multiple line "content" part is screwing json_decode(). Is there a way in PHP via RegEx or some nifty PHP that will strip out these items from the JSON output so that I can properly use json_decode() to retrieve n array that my code can then manipulate and deal with in the way I need it to?

Here is the error message given by JSONLint (and one other validator):

Parse error on line 263:
...         "content": "[TopNavigation_App
-----------------------^
Expecting 'STRING', 'NUMBER', 'NULL', 'TRUE', 'FALSE', '{', '['

I'm gathering that "[ is causing invalid JSON, but I'm no expert on Json.

I've attempted the following to no avail:

str_replace() on the lines in its original format
trim() and str_replace
trim()
str_replace removing the line breaks then attempting to str_replace() the culprit parts of the JSON.

Can you better describe what is invalid about it? Can you show us the code that json_decode() is failing on? — Brad
– Brad, Commented Aug 16, 2012 at 20:54
Added the JSONLint error message. However, the output isn't straight "content": "[, its on multiple lines as shown above. — Jeremy Dentel
– Jeremy Dentel, Commented Aug 16, 2012 at 21:01

Robert · Accepted Answer · 2012-08-16 21:06:29Z

2

Is the JSON coming in as a String? Its not the "[" but the new line in front of the square bracket that's throwing things off. If so you probably can use str_replace like:

$jsonStr = str_replace("\n", '', str_replace("\r", '', $jsonStr) );

But then its difficult to know for sure without more code.

answered Aug 16, 2012 at 21:06

Robert

2,44621 silver badges12 bronze badges

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

3 Comments

Jeremy Dentel Over a year ago

The full JSON document is over 3,330 lines of repeating json (in the structure above with nested JSON arrays inside of "items". It is being thrown on the line breaks of [TopNavigation_Appliances] every time (this happens upwards of 20 times).

Brad Over a year ago

@JeremyDentel, Did you try Robert's suggestion?

Jeremy Dentel Over a year ago

Yes. I'm still getting a null response from json_decode(). Is there a way to get a more exact error from PHP?

Jeremy Dentel · Accepted Answer · 2012-08-17 21:05:33Z

0

I solved this by using the following RegEx to strip out whitespace & then the offending "espots" from above:

preg_replace('#, "espots":.*?}\s]#', '', preg_replace('/\s+/', ' ',$nearlyThere[0]));

Then minor str_replaces on random invalid strings that weren't in the original post to get it 100% valid.

answered Aug 17, 2012 at 21:05

Jeremy Dentel

8472 gold badges9 silver badges19 bronze badges

Collectives™ on Stack Overflow

How to strip invalid characters from JSON output in PHP

2 Answers 2

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related