Regex to create two capture groups, where the second captures multiple times

Question

My test string is

thread_id=1152236, geo_locality.nomv="Seattle|||San Francisco|||Chicago", user_reference_count=0

Is it possible to have one regex to have two capture groups, where the second capture group will capture multiple times?

I want the first capture group to capture geo_locality (without hardcoding) and the second capture group to capture Seattle , San Francisco , and Chicago .

The closest I got was

(?<key>\w+)\.nomv="(?<val>.+?)(?=\|\|\||")

https://regex101.com/r/wmxg4x/1

Except the val capturing group also needs to capture the other cities.

Answer 1

The answer depends on the regex flavor at play.

1. Using \\G to continue at the end of the previous match with a PCRE

    (?<key>\\w+)\\.nomv="|(?!^)(?<=\\G)(?<val>.+?)(?:\\|\\|\\||") 

   Demo

   The \\G anchor can be a bit arcane and truly magical at the same time.

Explanation:

• (?<key>\\w+)\\.nomv="| the key and literal in the 1st alternation act as a start anchor

• The \\G anchor asserts the position at the end of the previous match or the start of the string for the first match.

  • to exclude the start of the string I've added (?!^) to prevent matches before nom=" ).
  • (?<=\\G) so, we can only continue if there was a match before
  • (?<val>.+?) captures each city block as required
  • (?:\\|\\|\\||") the non-capturing group is simply used to move the cursor forward

---

2. Using Captures with .NET

    (?<_KEY_1>\\w+)\\.nomv="(?:(?<_VAL_1>.+?)(?:\\|\\|\\||"))* 

   Demo

   This is not a real challenge for .NET. Just add a group around and a quantifier, and let (?<val>) match multiple times. Then, get the values from the Captures .

Answer 2

You could use an alternation with if supported a positive lookbehind (?<=

(?<_KEY_1>\\w+)(?=\\.nomv=")|(?<_VAL_1>(?<=\\.nomv=")[A-Za-z ]+|(?<=\\|\\|\\|)[A-Za-z ]+)

Explanation

• (?<_KEY_1> Named capture group
  • \\w+ Match one or more times a word character
• ) Close named capture group
• (?=\\.nomv=") Positive lookahead that asserts that what follows is .nomv="
• | Or
• (?<_VAL_1> Named capture group
  • (?<=\\.nomv=") Positive lookbehind that assert that what is on the left is .nomv=
  • [A-Za-z ]+ Match an uppercase, lowercase or whitespace
  • | Or
  • (?<=\\|\\|\\|) Positive lookbehind that assert that what is on the left is |||
  • [A-Za-z ]+ Match an uppercase, lowercase or whitespace
• ) Close named capture group

Answer 3

Try this pattern: (?<_KEY_1>\\w+)\\.nomv="(?<_VAL_1>(.+?\\|\\|\\|)+.+)" .

One thing to note is that inside capturing group _VAL_1 there can be some more capturing groups, but _VAL_1 is what you need.

See DEMO.