Extract a substring from value of key-value pair using regex - regex

Question

I have a string in log and I want to mask values based on regex.
For example:
"email":"testEmail#test.com", "phone":"1111111111", "text":"sample text may contain email testEmail#test.com as well"
The regex should mask
email value - both inside the string after "email" and "text"
phone number
Desired output:
"email":"*****", "phone":"*****", "text":"sample text may contain email ***** as well"
What I have been able to do is to mask email and phone individually but not the email id present inside the string after "text".
Regex developed so far:
(?<=\"(?:email|phone)\"[:])(\")([^\"]*)(\")
https://regex101.com/r/UvDIjI/2/

Answer 1

As you are not matching an email address in the first part by matching not a double quote, you could match the email address in the text by also not matching a double quote.
One way to do this could be to get the matches using lookarounds and an alternation. Then replace the matches with *****
Note that you don't have to escape the double quote and the colon could be written without using the character class.
(?<="(?:phone|email)":")[^"]+(?=")|[^#"\s]+#[^#"\s]+
Explanation
(?<="(?:phone|email)":") Assert what is on the left is either "phone":" or "email":"
[^"]+(?=") Match not a double quote and make sure that there is one at the end
| Or
[^#"\s]+#[^#"\s]+ Match an email like pattern by making use of a negated character class matching not a double quote or #
See the regex demo

Answer 2

Your current RegEx is trying to accomplish too much in a single take. You'd be better off splitting the conditions and dealing with them separately. I'll assume that the input will always follow the structure of your example, no edge cases:
Emails:
\w+#.+?(?="|\s) - In emails, every character preceded by # is always a word character, so using \w+# is enough to capture the first half of the email. As for the second half, I used a wildcard (.) with a lazy quantifier (+?) to stop the capture as soon as possible and combined it with a positive lookahead that checks for double quotes or whitespaces ((?="|\s)) so to capture both the e-mails inside "email" and "text" properties. Lookarounds are zero-length assertions and thus they don't get captured.
Phone number:
(?<="phone":")\d+ - Here I just use the prefix "phone":" in a lookbehind and then capture only digits \d+.
Combine both conditions and you have your RegEx: \w+#.+?(?="|\s)|(?<="phone":")\d+.
Regex101: https://regex101.com/r/UvDIjI/3

Answer 3

Meta Sequence Word Boundary \b & Alternation |
The input string pattern has either quotes or spaces wrapped around the targets which both are considered non-words. So this: "\bemailPattern\b" and this: space\bemailPattern\bspace are matches. The alternation gives one line the power of two lines. Search for emailPattern OR phonePattern.
/(\b\w+?#\w+?\.\w+?\b|[0-9]{10})/g;
(Word boundary (a non-word on the left) \b
One or more word characters \w+?
Literal #
One or more word characters \w+?
Escaped literal .
One or more word characters \w+?
Word boundary (a non-word on the right) \b
OR |
10 consecutive numbers [0-9]{10} )
global flag continues search after first match.
Demo
let str = `"email":"testEmail#test.com", "phone":"1111111111", "text":"sample text may contain email testEmail#test.com as well"`;
const rgx = /(\b\w+?#\w+?\.\w+?\b|[0-9]{10})/g;
let res = str.replace(rgx, '*****');
console.log(res);