Home / ubuntu / Questions / 1249777
Accepted
user1046658
Asked: 2020-06-13 14:03:44 +0800 CST

What does grouping in extended regular expressions mean?

  • 772

I read that grouping treats an expression as a single unit. What does that mean?

Furthermore, what does grep "\(grouping\)" file.txt or grep -E "(grouping)" file.txt do? I also read that expressions are grouped with parenthesis. Again, what is grouping expressions?

When I run the command, it highlights grouping in the file. How is grep "\(grouping\)" different from grep grouping?

command-line regex grep

1 Answers

  1. Best Answer

    Essentially, it causes whatever is inside the parentheses to be treated as a single atom. This is useful if you want to apply a quantifier for example. Compare:

    $ printf 'groupinggrouping' | grep -Eo "grouping*"
groupingg

    (where * is only applied to the g) with

    $ printf 'groupinggrouping' | grep -Eo "(grouping)*"
groupinggrouping

    where * is applied to the whole subpattern grouping.

    In most (all?) regex dialects, (grouping) also captures the matched text into an indexed capture group, allowing it to subsequently be backreferenced. That doesn't have too many applications in grep - the most obvious one is for detecting repeated elements ex.

    $ printf 'ba\nbb\nbc\n' | grep '\(.\)\1'
bb

    matches any single character that is followed by the same character. In the context of pattern substitution (in sed for example, rather than grep), the captured group may also be referenced in the replacement text.

    There are other variants in more expressive dialects - such as Perl's (?:grouping) non-capturing groups.

    For further information see for example

    • 3