Convert the character codes inside CDATA instead of the whole string-CodePudding

var str = '&<![CDATA[&]]><![CDATA[&]]>&';

In the above string, I just want to convert only & inside the CDATA not the all &.

Expected Output: &<![CDATA[&]]><![CDATA[&]]>&

I tried below regular expression

str.trim().replace(/^(\/\/\s*)?<!\[CDATA\[|(\/&\/\s*)?\]\]>$/g, '&');

But above code is not working as expected. I am not good in regular expressions. I gone through different answers given in Stackoverflow. But, not able to find the better way to achieve the fix. Could you please guide me.

CodePudding user response：

For this particular string you can apply /(?<=CDATA\[)[&a-z;] (?=]])/g

You can use positive lookbehind and lookahead:

(?<=CDATA\[) is a positive lookbehind. Searches everything after CDATA[
(?=]]) is a positive lookahead. Searches everything before ]]
[&a-z;] matches some text containing lowercase letters, & and ;

If I've got your idea correctly, it would be better to use XML parsers to manipulate a document.

Here you can find a sample js code.

regex101.com

CodePudding user response：

If you want to replace any & in CDATA, regardless of what comes before and after (within CDATA):

str.trim().replace(/<!\[CDATA\[.*?\]\]>/g, m => m.replace('&amp;', '&'));

results in

"&amp;<![CDATA[&]]><![CDATA[&]]>&amp;"

This first matches CDATA sections and replaces them with the result of a function, the function replaces all & with &;

Because that function is only applied on CDATA sections, &s outside of CDATA will not be changed.

Example with more characters in CDATA:

var str = '&amp;<![CDATA[Oh look at this: &amp; Haha!]]>&amp;';
str.trim().replace(/<!\[CDATA\[.*?\]\]>/g, m => m.replace('&amp;', '&'));

result:

"&amp;<![CDATA[Oh look at this: & Haha!]]>&amp;

CodePudding user response：

If you have control over the data received it is better to fix the data upstream. If not, you can use nested replaces:

outer replace identifies the <![CDATA[...]]>
inner replace & inside CDATA

Both use the g flag to replace multiple time.

[
  '&amp;<![CDATA[&amp;]]><![CDATA[&amp;]]>&amp;',
  '&amp;<![CDATA[this &amp; that]]>&amp;'
].forEach(str => {
  let result = str.replace(/<!\[CDATA\[[^\]]*\]\]>/, m => m.replace(/&amp;/g, '&'));
  console.log(result);
});

Output:

&amp;<![CDATA[&]]><![CDATA[&amp;]]>&amp;
&amp;<![CDATA[this & that]]>&amp;