How can I specify a regex in a zgrep/zcat command?

Question

I want to find in a list of words, every words with a least 3 times the same letter in it. To achieve that I did .*(\w).*\1.*\1.*\1.* and you can test it here http://www.regexplanet.com/advanced/java/index.html but I don't know how to put it in my zgrep command.

How can I insert this regex in a zgrep command ?

Hey you already helped me earlier ! I tried but I have `invalid back reference` — J.erome, Oct 16 '19 at 09:21
Yes, you also need to create a capturing group. Try `zgrep '\([[:alnum:]_]\).*\1.*\1.*\1' a.gz` — Wiktor Stribiżew, Oct 16 '19 at 09:24

score 1 · Accepted Answer · answered Oct 16 '19 at 09:32

1

A couple of notes:

You do not need to match start and end of a line with .* since partial matches are allowed
\w matches letters, digits and underscores in NFA regex patterns, in POSIX, it is safer to use [[:alnum:]_]
To form a capturing group in a POSIX BRE pattern use escaped parentheses, \(...\).

Thus, use

zgrep '\([[:alnum:]_]\).*\1.*\1.*\1' a.gz

Or, contract it a bit since it looks a bit redundant with three consecutive .*\1 subpatterns:

zgrep '\([[:alnum:]_]\)\(.*\1\)\{3\}' a.gz

answered Oct 16 '19 at 09:32

Wiktor Stribiżew

607,720
39
448
563

Is it a medium to contact you ? I have more questions about regex and zgrep but I don't want to spam creating a lot of posts – J.erome Oct 16 '19 at 09:39
@J.erome Please join the [chat](https://chat.stackoverflow.com/rooms/200941/room-for-wiktor-stribizew-and-j-erome). – Wiktor Stribiżew Oct 16 '19 at 09:44
So, to match 3 or more times, use `\{3,\}` instead of `\{3\}`. – Wiktor Stribiżew Oct 16 '19 at 09:56

How can I specify a regex in a zgrep/zcat command?

1 Answers1