"sed" command to remove a line that match an exact string on first word

Question

I'm working in a small bash script. I need to remove an entire line that match an exact word in only the first word of the line.

That's how the text file looks:

John is a dumb
Maria is awesome
Toni and Maria are funny

I want to remove the just the second line.

Right now, I can match the exact word "maria" but it removes the third line too:

sed -i "/\b\(maria\)\b/d" file.txt

How to specify just the first word?

Thanks!

Anchor your pattern to the start of the line with `^`. [The Anchor Characters](http://www.grymoire.com/Unix/Regular.html#uh-2) — Etan Reisner, Oct 29 '15 at 16:13
you can try `sed -i "/^maria\b/I d" file.txt` ..... `^` is start line and `I` for case insensitive — Jose Ricardo Bustos M., Oct 29 '15 at 16:15
Why is this question downvoted? It has a precise description of the goal, a sample dataset, and an own approach to reach the goal. — Rambo Ramon, Oct 29 '15 at 16:17
@RamboRamon thank you very much, I was really upset with that! — Dani Polo, Oct 29 '15 at 16:36

miken32 · Accepted Answer · 2015-10-29T16:41:12.967

6

Currently you are looking for "maria" surrounded by a word boundary (represented by \b.) Instead look for "maria" preceded by the start of line (represented by ^.) Note I've also removed unnecessary parentheses and added the /I flag, which will make the search case-insensitive. Your original wouldn't have matched "Maria".

sed -i "/^maria\b/Id" file.txt

Edit: fixed case insensitive flag to I instead of i!

edited Oct 29 '15 at 16:41

answered Oct 29 '15 at 16:35

miken32

42,008
16
111
154

It doens't work in ksh? If text has line as 'maria-d' then? doesn't work – rohansr002 Jan 30 '20 at 13:01
ksh is a shell. sed is a program. They aren’t really related. I addressed the sample data in the question, but I’m sure in most languages the word boundary assertion matches hyphens. – miken32 Jan 30 '20 at 14:30

"sed" command to remove a line that match an exact string on first word

1 Answers1

Linked