How to avoid capturing groups if the captured match is empty?

user_7010473 0
Viewed 80

I would like to prepend the word "custom" to a list of host-names whose subdomains can be separated by some separator.

Examples:

news.google.com   -> custom.news.google.com
news/google/com   -> custom.news.google.com

dev.maps.yahoo.fr -> custom.dev.maps.yahoo.fr
dev/maps/yahoo/fr -> custom/dev/maps/yahoo/fr

These strings appear inside a document with more content, so I am trying to solve this problem using regular expressions and JavaScript's string replace function.

The list of hostnames and separators is predefined and known in advance. For the sake of this example, I only showed 2 hostnames (news.google.com and dev.maps.yahoo.com) and 2 separators (. and /), but there are more.

The separator within a single string will always be the same, i.e. there won't be cases like dev/maps.yahoo/fr.

I want to be consistent and use the correct separator when prepending "custom".

I built this long regular expression:

const myRegex = /news\.google\.com|news\/google\/com|dev\.maps\.yahoo\.fr|dev\/maps\/yahoo\/fr/

(For readability purposes, this is the expression:

/news\.google\.com/ OR /news\/google\/com/ OR /dev\.maps\.yahoo\.fr/ OR /dev\/maps\/yahoo\/fr/ )

(Note: It is important to emphasize that the list of hostnames is predefined and well known in advance, that's why I am 'hardcoding' the hostnames and not using tokens such as \w+ or \S+. For example, I might want to replace news.google.com, but leave news2.google.com intact).

However, I am not sure how to capture the separator (whether ., /, or any other separator). I tried using capture groups like this:

const myRegex = /news(\.)google\.com|news(\/)google\/com|dev(\.)maps\.yahoo\.fr|dev(\/)maps\/yahoo\/fr/

However, by doing this, I am creating 4 capture groups, and there's only one separator (and this is just a simple example). 3 of the capture groups will be empty, and one of them will contain the separator. How can I know which capture group is it?

Ideally, I would like something like this:

const myString = 'I navigated to news.google.com'; // For example
const myCustomString = myString.replace(
  myRegex,
  (match, <SEPARATOR_WRONG>) => `custom${SEPARATOR_WRONG}${match}`,
);

console.log(myCustomString); 
// will log 'I navigated to custom.news.google.com'

Is there a way to skip captured groups if they are empty?

javascript regex
2 Answers

Use \1 to refer to the separator captured in the first (\.|\/) group so we don't have to keep writing it over and over.

const text = `I navigated to news.google.com
I navigated to news/google/com
I navigated to dev.maps.yahoo.fr
I navigated to dev/maps/yahoo/fr`;

const re = /\w+(\.|\/)(\w+\1)?(google|yahoo)\1\w+/g;
console.log(text.replace(re, (url, separator) => `custom${separator}${url}`));

Here's an alternate solution given the new requirement described in the comments:

const text = `I navigated to news.google.com
I navigated to news/google/com
I navigated to dev.maps.yahoo.fr
I navigated to dev/maps/yahoo/fr`;

const re = /(news|dev)(\.|\/)(google|maps)\2(com|yahoo)(\2fr)?/g;

console.log(text.replace(re, (url, prefix, separator) => `custom${separator}${url}`));

Yet another alternate solution:

const text = `I navigated to news.google.com
I navigated to news/google/com
I navigated to dev.maps.yahoo.fr
I navigated to dev/maps/yahoo/fr`;

const re = /news(\.)google\.com|news(\/)google\/com|dev(\.)maps\.yahoo\.fr|dev(\/)maps\/yahoo\/fr/g;

console.log(text.replace(re, url => 'custom' + url.match(/\.|\//)[0] + url));

user_133782470

solution that I believe is acceptable to you is to add separator finding logic among the capture groups in the callback

const myCustomString = myString.replace(
  myRegex,
  (match, ...rest) => {
     const sep = rest.slice(0, -2) // last two args are index and full match
      .find(sep => !!sep) // first truthy capture group contains a separator
      return `custom${sep}${match}`},
);
user_110788570
Related