Text manipulation: Extract everything inside brackets

Question

I have a text file where every line is in a similar format to this:

%#&#%#    [Dinero]    / Money / 
!#@%$@    [Dia]     / Day /
$%&$^#@ [Perro]   / Dog / 

I am looking to extract the words inside the brackets, ie. Ola, Dinero, Perro, etc, and save it all to a new text file line by line. Essentially, I am looking to omit/delete/erase all words, letters, special characters, and anything else outside the brackets, including the brackets themselves.

Solution 1

with the help of awk,

$ awk -F'[][]' '{print $2}' < input
Dinero
Dia
Perro

Using grep,

grep -oP '\[\K[^\]]+' input

\K means that use look around regex advanced feature. More precisely, it's a positive look-behind assertion

if you lack the -P option, you can do this with perl:

perl -lne '/\[\K[^\]]+/ and print $&' input

use -i option to edit file in place.

Or simply you can use cut as suggested by @juliepelletier,

cut -d"[" -f2 < input | cut -d"]" -f1

Solution 2

sed 's/^.*\[//;s/\].*$//' /path/to/input > /path/to/output

Comments

[Julie Pelletier]

For the sake of simplicity, I was thinking of answering: cat input|cut -d\[ -f2|cut -d\] -f1

[Rahul]

@juliepelletier yes, updated your suggestion in answer. Thanks

[VLAZ]

huh, I've not actually used the \K flag. seems useful. Given the requirements, I'd have simply done grep -o -e '\[.*\]' file.txt | sed -e 's/\[//' -e 's/\]//' to find and extract all words with the brackets surrounding them and then strip away the brackets. And yes, that's the slightly verbose variant - could also be done more concisely with the same effect. It just better describes the flow I'd have put the data through. I keep meaning to pick up awk but I can never find the time.

[Matthew Finlay]

I think you need <input, not > input

[Rahul]

@MatthewFinlay I have not used > input anywhere in my post.

[Matthew Finlay]

@Rahul $ awk -F'[][]' '{print $2}' > input

[Rahul]

@MatthewFinlay Ah, silly typo. Anyway thanks for pointing out.