regex - matching between a literal string and a quotation mark

Question

I'm terrible at Regex and would greatly appreciate any help with this issue, which I think will be newb stuff for anyone familiar.

I'm getting a response like this from a REST call

    {"responseData":{"translatedText":"Ciao mondo"},"responseDetails":"","responseStatus":200,"matches":[{"id":"424913311","segment":"Hello World","translation":"Ciao mondo","quality":"74","reference":"","usage-count":50,"subject":"All","created-by":"","last-updated-by":null,"create-date":"2011-12-29 19:14:22","last-update-date":"2011-12-29 19:14:22","match":1},{"id":"0","segment":"Hello World","translation":"Ciao a tutti","quality":"70","reference":"Machine Translation provided by Google, Microsoft, Worldlingo or the MyMemory customized engine.","usage-count":1,"subject":"All","created-by":"MT!","last-updated-by":null,"create-date":"2012-05-14","last-update-date":"2012-05-14","match":0.85}]}

All I need is the 'Ciao mondo' in between those quotations. I was hoping with Java's Split feature I could do this but unfortunately it doesn't allow two separate delimiters as then I could have specified the text before the translation.

To simplify, what I'm stuck with is the regex to gather whatever is inbetween translatedText":" and the next "

I'd be very grateful for any help

[You're asking an XY question.](http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem) Regex is the wrong tool for the job. Are you sure you wouldn't rather parse the JSON? [See this](http://meta.stackexchange.com/a/66378/182647). — Li-aung Yip, May 14 '12 at 03:09
And this [isn't even the first time](http://stackoverflow.com/questions/9832954/regex-issue-scraping-youtube) you've tried to apply regexes to problems better solved by more specific tools. As per my answer to your earlier question, trying to invent your own mini-parser for a language where parsers *already exist* is a losing game. Use a JSON parser. — Li-aung Yip, May 14 '12 at 03:25
Wow. This actually worked perfectly and was very easy to implement. As this wasn't a proper 'answer', do you know how I finalize the question? — CitizenSmif, May 14 '12 at 03:34
You don't *have* to accept any answer. You could accept DasBlinkenlight's answer (which is a perfect answer to your original question.) — Li-aung Yip, May 14 '12 at 03:46

score 3 · Accepted Answer · answered May 14 '12 at 03:25

You can use \"translatedText\":\"([^\"]*)\" expression to capture the match.

The expression meaning is as follows: find quoted translatedText followed by a colon and an opening quote. Then match every character before the following quote, and capture the result in a capturing group.

String s = " {\"responseData\":{\"translatedText\":\"Ciao mondo\"},\"responseDetails\":\"\",\"responseStatus\":200,\"matches\":[{\"id\":\"424913311\",\"segment\":\"Hello World\",\"translation\":\"Ciao mondo\",\"quality\":\"74\",\"reference\":\"\",\"usage-count\":50,\"subject\":\"All\",\"created-by\":\"\",\"last-updated-by\":null,\"create-date\":\"2011-12-29 19:14:22\",\"last-update-date\":\"2011-12-29 19:14:22\",\"match\":1},{\"id\":\"0\",\"segment\":\"Hello World\",\"translation\":\"Ciao a tutti\",\"quality\":\"70\",\"reference\":\"Machine Translation provided by Google, Microsoft, Worldlingo or the MyMemory customized engine.\",\"usage-count\":1,\"subject\":\"All\",\"created-by\":\"MT!\",\"last-updated-by\":null,\"create-date\":\"2012-05-14\",\"last-update-date\":\"2012-05-14\",\"match\":0.85}]}";
System.out.println(s);
Pattern p = Pattern.compile("\"translatedText\":\"([^\"]*)\"");
Matcher m = p.matcher(s);
if (!m.find()) return;
System.out.println(m.group(1));

This fragment prints Ciao mondo.

Untitled · Answer 2 · 2012-05-14T03:26:28.360

use look-ahead and look-behind to gather strings inside quotations: (?<=[,.{}:]\").*?(?=\")

class Test
{
    public static void main(String[] args)
    {
        Scanner scanner = new Scanner(System.in);
        String in = scanner.nextLine();

        Matcher matcher = Pattern.compile("(?<=[,.{}:]\\\").*?(?=\\\")").matcher(in);

        while(matcher.find())
            System.out.println(matcher.group());
    }
}

Pradeep · Answer 3 · 2012-05-14T03:34:44.987

0

Try this regular expression -

^.*translatedText":"([^"]*)"},"responseDetails".*$

The matching group will contain the text Ciao mondo.

This assumes that translatedText and responseDetails will always occur in the positions specified in your sample.

edited May 14 '12 at 03:34

answered May 14 '12 at 03:25

Pradeep

148
1
9

It also assumes that the `Translated Text` will only contain `[a-zA-Z\s]`. What if the translated text is `¡Santo cielo! ¿Es que una vaca?`? Or possibly the same thing in Russian - `Святая корова! Это корова?` – Li-aung Yip May 14 '12 at 03:29
Then Houston we have a problem! I like dasblinkenlight's solution. I am revising mine to accept everything that's not a double quote. – Pradeep May 14 '12 at 03:33
@Li-aungYip That's a poor translation - the `¡Santo cielo!` is commonly translated as `Боже мой!` – Sergey Kalinichenko May 14 '12 at 03:39
@dasblinkenlight: Google Translate gave that to me (source text: "Holy cow! Is that a cow?") Perhaps you should rate the Google Translation for improvement. ;) – Li-aung Yip May 14 '12 at 03:59

regex - matching between a literal string and a quotation mark

3 Answers3