Crumb
c/linux · by Corsair@programming.dev · 1y

regex and awk...

cross-posted from: https://programming.dev/post/31833654

Hi,

I would like to found a regex match in a stdout

stdout

 /dev/loop0: [2081]:64 (/a/path/to/afile.dat)

I would like to match

/dev\/loop\d/

and return /dev/loop0

but the \d seem not working with awk ... ?

How to achieve this ? ( awk is not mandatory )

20 points · 9 comments · view on lemmy.world

9 Comments

ik5pvx@lemmy.world · 6 pts · 1y (1 reply)

Use [0-9] to match the number

bizdelnick@lemmy.ml · 5 pts · 1y

Or, alternatively, [[:digit:]], and dont' forget to add a quntifier + to match multiple digits. See documentaion for details.

awk '/^\/dev\/loop[[:digit:]]+/{print}'
mmmm@sopuli.xyz · 4 pts · 1y (4 replies)

Not sure if I'm understanding, but can't you just pipe the whole thing to awk and capture the first field? Like

echo "/dev/loop0: [2081]:64 (/a/path/to/afile.dat)" | awk -F: '{print $1}'

Which would print

/dev/loop0

lungdart@lemmy.ca · 0 pts · 1y (3 replies)

That would also print the colon

Edit: missed the separator token. Sorry guys

manxu@piefed.social · 2 pts · 1y

The field separator is declared to be the colon, with -F:, so the fields end and start at colons.

mmmm@sopuli.xyz · 2 pts · 1y

No, because we're telling to use : as a separator with the -F flag

learnbyexample@programming.dev · 1 pts · 1y

Why would it print the colon?

learnbyexample@programming.dev · 4 pts · 1y (1 reply)

Regex syntax and features vary between implementations. \d isn't supported by BRE/ERE flavors.

GNU grep supports PCRE, so you can use grep -oP '/dev/loop\d' or grep -o '/dev/loop[0-9]' if you are matching only one digit character.

4am@lemm.ee · 4 pts · 1y

I wish there was one single unifying regex standard.

(obligatory xkcd in 3..2..)