Extract only whole word using grep

Question

I've got a big text file. I need to extract all the lines which contains the exact word "DUSP1". Here an example of the lines:

9606    ENSP00000239223 DUSP1   BLAST
9606    ENSP00000239223 DUSP1-001 Ensembl

I want to retrieve the first line but not the second one.

I tried several commands as:

grep -E "^DUSP1"
grep '\<DUSP1\>'
grep '^DUSP1$'
grep -w DUSP1

But none of them seem to work. Which option should I use?

Metatarsus

How exactly is the "exact word" defined? And your 3rd example would only find lines with only the word "DUSP1" ... So you want lines with "^DUSP1[[:space:]]+" ? – Claribelclarice

Could you provide sample file content. The 2nd, 3rd, 4th commands works for me. – Drogheda

Answer 1

The problem you are facing is that a dash (-) is considered by grep as a word delimiter.

You should try this command :

grep '\sDUSP1\s' file

to ensure that there's spaces around your word.
Or use words boundaries :

grep '\bDUSP1\b' file

Backplate

Answer 2

If you want to grep exactly the whole word, you can use word boundaries like this:

grep '\bDUSP1\b'

This matches for the exact word at the beginning and at the end.

Knight

This should be the accepted answer, there are not always spaces before and after (what about when it's the last word?). – Lingonberry

this is great. It also matches setting=DUSP1 and my/folder/to/DUSP1, but not DUSP123 – Sacramentalism

I had to use double quotes for the Windows version of GNU grep. Single quotes did not work. – Hoch

Yep, this should be the answer with the big green checkmark. :D – Cardsharp

this still greps DUSP1-001 for me – Cheerless

no does not work - still greps words with DUSP1-xxx – Dumfries

Answer 3

adding to what sputpick said, it could either be that or:

grep '\sDUSP1$' file 

if the DUSP1 is the end of the line.

Nettlesome