How to delete the n-th word from standard input?

Question

I want to create a script that deletes the N^th word from the standard input, for a given N. For example, for this input:

One two three four, five
six seven eight, nine

If we ask to delete the 8th word, it should delete the eight,. For my purposes, a word is any sequence of non-space characters.

One two three four, five
six seven  nine

Is there some clever one-liner that can accomplish this using standard command line utilities? Currently I have a fairly long script to do this, but it feels like overkill.

Does your real file has only two lines? I believe not... is better to update the question with more data. — George Vasiliou, Sep 02 '22 at 18:13
Does this answer your question? Using sed to replace nth occurrence of a word — G-Man Says 'Reinstate Monica', Sep 06 '22 at 00:54
Related: How to add text before the Nth occurrence of a text using sed only?,  sed or awk: replace only the n-th occurrence of a string,  BSD sed: Replace only the Nth occurrence of a pattern and Print everything after nth delimiter — G-Man Says 'Reinstate Monica', Sep 06 '22 at 00:54

Stéphane Chazelas · Accepted Answer · 2022-09-03T05:57:37.513

3

With perl:

$ perl -pe 's/\S+/++$c == 8 ? "" : $&/ge' <your-file
One two three four, five
six seven  nine

$ perl -pse 's/\S+/--$n ? $& : ""/ge' -- -n=8 <your-file
One two three four, five
six seven  nine

Or optimising a bit and not perform the substitutions once the n^th word has been found:

perl -pse 's/\S+/--$n ? $& : ""/ge if $n > 0' -- -n=8 <your-file

edited Sep 03 '22 at 05:57

answered Sep 02 '22 at 18:16

Stéphane Chazelas

544,893

score 2 · Answer 2 · answered Sep 02 '22 at 17:35

2

Using any awk and without reading all of the input into memory at once:

$ awk -v t=8 '{p=n; n+=NF} (n>t) && !f++{$(t-p)=""} 1' file
One two three four, five
six seven  nine

answered Sep 02 '22 at 17:35

Ed Morton

31,617

3

Would be worth noting that it may affect the spacing in the line with the word to remove as sequences of one or more whitespace characters are turned into one space and the leading and trailing whitespace are removed. – Stéphane Chazelas Sep 02 '22 at 18:20

score 1 · Answer 3 · edited Sep 02 '22 at 16:45

1

Using GNU sed

$ sed -Ez 's/(([^ \n]*( |\n)){7})[^ ]*/\1/' input_file
One two three four, five
six seven  nine

edited Sep 02 '22 at 16:45

hugomg

5,747
4
39
54

answered Sep 02 '22 at 16:32

sseLtaH

2,786

2

Hmm... what does this do that the naive sed -z 's/[^[:space:]]\{1,\}//8' doesn't? – steeldriver Sep 02 '22 at 17:01
5

Since -z for reading the whole file into memory requires GNU sed you could also take advantage of other GNU-isms and abbreviate that to sed -Ez 's/\S+//8' – Ed Morton Sep 02 '22 at 17:41

QuartzCristal · Answer 4 · 2022-09-03T13:19:14.927

1

A simple solution with awk:

awk 'n>0 && n<=NF {$n=""} {n-=NF} 1' n=8 infile

If you need to re-adjust the white spaces (as one field has been removed, two consecutive FS appear).

$ awk 'n>0 && n<=NF {$n="";gsub(/[ \t]+/, " ")} {n-=NF} 1' n=8 infile
One two three four, five
six seven nine

edited Sep 03 '22 at 13:19

answered Sep 03 '22 at 05:11

QuartzCristal

2,023

score 0 · Answer 5 · answered Sep 08 '22 at 21:01

This solution uses GNU sed's -z feature, which reads the whole file as a "single line"

sed -Ez 's/\S+//8'

Alternatively, to also remove the spaces following that word

sed -Ez 's/\S+ *//8'

Credit goes to Ed Morton, who posted this as a comment to another answer.

How to delete the n-th word from standard input?

5 Answers5

Linked