sed: replace from the last occurrence of "-" till EOL

Question

My problem

I have a file with items, one per line. The item contains at least one dash, and I would like to remove the last dash and the words that follow. For example:

item asdkalqndla-asdnkfsv-324we-blueray
item asda-vbght564e-dfg-redapple
item gefdsc-fgy-543-5trr-floatingvanilla

Should give:

item asdkalqndla-asdnkfsv-324we
item asda-vbght564e-dfg
item gefdsc-fgy-543-5trr

What have I tried

sed 's/\-.*$//' lines.txt

Which gives

item asdkalqndla
item asda
item gefdsc

Because the regex is greedy, and consumes everything from the first dash onwards.

My question

How can I remove all characters from the last - in a string till EOL?

Possible duplicate of [Display all fields except the last](https://stackoverflow.com/questions/13857836/display-all-fields-except-the-last) — Benjamin W., Nov 12 '17 at 18:21

score 3 · Answer 1 · answered Nov 12 '17 at 17:17

3

With awk:

$ awk '{sub(/-[^-]*$/,""); print}' file
item asdkalqndla-asdnkfsv-324we
item asda-vbght564e-dfg
item gefdsc-fgy-543-5trr

answered Nov 12 '17 at 17:17

dawg

98,345
23
131
206

score 2 · Answer 2 · edited Nov 12 '17 at 17:11

2

Simple negation using ^\- solved the problem:

$ sed 's/\-[^\-]*$//' lines.txt
item asdkalqndla-asdnkfsv-324we
item asda-vbght564e-dfg
item gefdsc-fgy-543-5trr

This way, sed replaces a dash followed by anything-but-dash till the end of the line.

edited Nov 12 '17 at 17:11

hek2mgl

152,036
28
249
266

answered Nov 12 '17 at 17:01

Adam Matan

128,757
147
397
562

You don't need to escape dashes. – SLePort Nov 12 '17 at 17:46

RavinderSingh13 · Answer 3 · 2017-11-12T18:09:58.230

2

Following simple awk may help you in same.

awk -F"-" 'NF{NF=(NF-1)} 1' OFS="-"   Input_file

Output will be as follows.

item asdkalqndla-asdnkfsv-324we
item asda-vbght564e-dfg
item gefdsc-fgy-543-5trr

Explanation: -F"-": Making field separator as - for each line for Input_file.

NF{NF=(NF-1)}: awk has out of the box NF variable which has number of fields for any line, since we don't text after last occurrence of - and field separator is -, by mentioning NF, we are checking if a line is NOT EMPTY. decrementing the number of field's value with 1 so that we will NOT get the last field.

1: awk works on method of condition then action so making condition part TRUE here by mentioning 1 here and not mentioning any action here so by default print of current line will happen.

OFS="-": Making OFS output field separator as "-" here so that - will come in output.

EDIT: Shorter versions of decrementing NF too as follows.

awk -F"-" 'NF{NF-=1} 1' OFS="-"  Input_file

OR

awk -F"-" 'NF{NF--} 1' OFS="-"  Input_file

edited Nov 12 '17 at 18:09

answered Nov 12 '17 at 17:39

RavinderSingh13

130,504
14
57
93

1

writing `NF--` is shorter. – Casimir et Hippolyte Nov 12 '17 at 17:54
@CasimiretHippolyte, then I could use `NF-=1` too I think :) – RavinderSingh13 Nov 12 '17 at 17:55
I don't think you can assign a negative value to NF. That's a problem. If the line has zero field, it produces an error. Conclusion it's better to write: `NF&&NF--` – Casimir et Hippolyte Nov 12 '17 at 17:56
@CasimiretHippolyte, mentioned it now, thank you for letting me know. – RavinderSingh13 Nov 12 '17 at 18:12

score 1 · Answer 4 · answered Nov 12 '17 at 22:52

1

This might work for you (GNU sed):

sed 's/\(.*\)-.*/\1/' file

Used greed to find the last - and replace the whole line with what came before it.

answered Nov 12 '17 at 22:52

potong

55,640
6
51
83

sed: replace from the last occurrence of "-" till EOL

My problem

What have I tried

My question

4 Answers4