How to make ripgrep match multiple rules in one line or use longest match principle？

[XhstormR]

This is the content of the data.txt:

www.yxlink.com/
www.china.com/
www.goodfood.com/

This is the content of the rule.txt:

(?-u:www.)
(?-u:www.yxlink.com/)
(?-u:www.china.com/)
(?-u:www.goodfood.com/)

This is the command:

rg -f rule.txt data.txt

This is the result after execution:

You can see that only the first rule was matched successfully, and all the following rules failed to match.

When I remove the first rule (?-u:www.), it can be matched successfully:

How to make ripgrep match multiple rules in one line or use longest match principle？

[BurntSushi]

The only way is to sort your rules file by the length of each rule in descending order.

[BurntSushi]

For further reading, please see my aho-corasick docs on match kind: https://docs.rs/aho-corasick/0.7.13/aho_corasick/enum.MatchKind.html --- It's probably the best I've written on the topic.

If you have a specialized task where all you need to do is search for a fixed set of non-regex strings, then it would be very simple to write a Rust program that uses the aho-corasick crate and enables leftmost longest match semantics. Then you wouldn't need to sort your rules, if that matters.

I do feel obligated to note though that your existing setup is interpreted your rules as regexes. In particular, the . is a special meta-character that matches any character, and you probably want to have it match a literal .. You can fix this by either escaping the . via \., or by passing the -F flag to ripgrep, which will treat the rules as literal non-regex patterns.

[XhstormR]

Thank you very much for your help!