Remove exact matching word using sed

Question

I want to remove word 'foggy' from the string. It fails. Why?

echo 'foggy light' | sed 's/\<foggy\>//g'

What version of Ubuntu? What's the output of sed --version? — muru, Feb 18 '18 at 11:39
@userunknown of course: GNU sed, installed by default on all versions of Ubuntu, has \< and \> for matching start-of-word and end-of-word. I suspect this is yet another Android question. OP has been know to post a number of questions about some Android app that provides Unix commands, while pretending it is Ubuntu OP is using. — muru, Feb 18 '18 at 12:24
For removing the word foggy in above scenario, you don't need start-of-word/end-of-word-markers. — user unknown, Feb 18 '18 at 12:33
Which it is not on any current version of Ubuntu: https://packages.ubuntu.com/search?keywords=sed, so yet again an off-topic question. — muru, Feb 18 '18 at 13:18

score 14 · Accepted Answer · edited Feb 18 '18 at 13:35

14

In the above string, you don't need start-of-word/end-of-word markers and you may use:

echo 'foggy light' | sed 's/foggy//g'

For the additional question in the comment:

Indeeed, my sed version

sed --version
sed (GNU sed) 4.2.2

supports the Syntax with \<...\>

echo 'foggy foggylight' | sed 's/\<foggy\>//g'
foggylight

If it doesn't work for you, report your sed version and read its manpage. For my sed, this syntax works too:

echo 'foggy foggylight' | sed 's/\bfoggy\b//g'
foggylight

\b can be memorized as boundary.

edited Feb 18 '18 at 13:35

muru

answered Feb 18 '18 at 12:18

user unknown

1

It's not < and > but \< and\>. See https://www.gnu.org/software/sed/manual/sed.html#regexp-extensions – muru Feb 18 '18 at 12:26
Ok. And how to drop just the word foggy alone, not touching foggylight? echo 'foggy foggylight' | sed 's/foggy//g' – Josef Klimuk Feb 18 '18 at 13:11
1

@JosefKlimuk: Added extension – user unknown Feb 18 '18 at 13:35

1 Answers1