How to remove last characters from a string in bash using sed?

Question

I would like to change all occurrences like "waaa", "wwwaaaaa", "wa" by trimming all the "a" in the end.

waaa > w
wwa > ww
waaaaa > w

I know how to find these strings in a text using regular expressions:

grep -nE "wa+" file.txt

I also know how to change one string in a bash script

#!/bin/bash
s1="wwwwaaaaaaaa"
s2=${s1%w*}
echo "$s1 --> $s2"

But I would like to use someting like

sed -E 's/wa+/ZZZ/' file.txt

where ZZZ is something that I need here.

sed -E 's/wa+/${$1%w*}/' file.txt

doesn't work.

Sample input:

Lorem ipsum wa waaaaaaa saaa dolor sit amet,
consectetur aw awwwwaaaa adipiscing elit

Desirable output:

Lorem ipsum w w saaa dolor sit amet,
consectetur aw awwww adipiscing elit

Using `gnu sed`, you can do: `sed -E 's/\bwa+/w/g' file` – anubhava Jan 04 '18 at 19:04 — anubhava, Jan 04 '18 at 19:04

PesaThe · Answer 1 · 2018-01-04T21:31:21.377

5

sed 's/waa*\b/w/g' file

sed 's/waa*\>/w/g' file

or using extended regex:

sed -r 's/wa+\b/w/g' file

\b and \> are word boundaries. For more information, see this regex tutorial: Word Boundaries.

edited Jan 04 '18 at 21:31

answered Jan 04 '18 at 17:53

PesaThe

1

@maro then you should mention you want this for multiple words. Your input implies you have one long string. – PesaThe Jan 04 '18 at 18:16
I think I wrote it in the question. – maro Jan 04 '18 at 18:22

score 0 · Answer 2 · answered Jan 04 '18 at 17:57

0

With bash, use extended pattern matching to remove all the "a" from the end of the string:

shopt -s extglob
for str in "waaa" "wwwaaaaa" "wa"; do
    new="${str%%+(a)}ZZZ"
    echo "$str => $new"
done

waaa => wZZZ
wwwaaaaa => wwwZZZ
wa => wZZZ

answered Jan 04 '18 at 17:57

glenn jackman

2 Answers2