nixCraft

Linux and Unix tutorials for new and seasoned sysadmin

How To Use grep Command In Linux /

UNIX
Author: Vivek Gite
Last updated: July 3, 2020

The grep command is used to search text. It searches the given file for lines
containing a match to the given strings or words. It is one of the most useful
commands on Linux and Unix-like system. Let us see how to use grep on a Linux
or Unix like system.

Did you know?

The name, “grep”, derives from the command used to perform a similar
operation, using the Unix/Linux text editor ed:
g/re/p

The grep utilities are a family that includes grep, egrep, and fgrep for searching
duties. For most uses, you need to use fgrep as it the fastest and only look into
strings and words. However, typing grep is easy. Hence, it is a personal choice.

grep command examples in Linux and Unix

Below is some standard grep command explained with examples to get you

1 of 12
started with grep on Linux, macOS, and Unix:

1. Search any line that contains the word in filename on Linux:

grep 'word' filename

2. Perform a case-insensitive search for the word ‘bar’ in Linux and Unix:
grep -i 'bar' file1

3. Look for all files in the current directory and in all of its subdirectories in Linux
for the word ‘httpd’:
grep -R 'httpd'

4. Search and display the total number of times that the string ‘nixcraft’ appears
in a file named frontpage.md:
grep -c 'nixcraft' frontpage.md

Let us see all commands and options in details.

Syntax
The syntax is as follows:

grep 'word' filename

fgrep 'word-to-search' file.txt
grep 'word' file1 file2 file3
grep 'string1 string2' filename
cat otherfile | grep 'something'
command | grep 'something'
command option1 | grep 'data'
grep --color 'data' fileName
grep [-options] pattern filename
fgrep [-options] words file

How do I use grep to search a file on Linux?

Search /etc/passwd file for boo user, enter:

2 of 12
 grep boo /etc/passwd

Sample outputs:

foo:x:1000:1000:boo,,,:/home/boo:/bin/ksh

We can use fgrep/grep to find all the lines of a file that contain a particular word.
For example, to list all the lines of a file named address.txt in the current directory
that contain the word “California”, run:

fgrep California address.txt

Please note that the above command also returns lines where “California” is part
of other words, such as “Californication” or “Californian”. Hence pass the -w

option with the grep/fgrep command to get only lines where “California” is
included as a whole word:

fgrep -w California address.txt

You can force grep to ignore word case i.e match boo, Boo, BOO and all other
combination with the -i option. For instance, type the following command:

grep -i "boo" /etc/passwd

3 of 12
The last grep -i "boo" /etc/passwd can run as follows using the cat command
too:

cat /etc/passwd | grep -i "boo"

How to use grep recursively

You can search recursively i.e. read all files under each directory for a string
“192.168.1.5”

$ grep -r "192.168.1.5" /etc/

OR

$ grep -R "192.168.1.5" /etc/

Sample outputs:

/etc/ppp/options:# ms-wins 192.168.1.50

4 of 12
 /etc/ppp/options:# ms-wins 192.168.1.51
/etc/NetworkManager/system-connections/Wired connection 1:
addresses1=192.168.1.5;24;192.168.1.2;

You will see result for 192.168.1.5 on a separate line preceded by the name of
the file (such as /etc/ppp/options) in which it was found. The inclusion of the file
names in the output data can be suppressed by using the -h option as follows:

$ grep -h -R "192.168.1.5" /etc/

OR

$ grep -hR "192.168.1.5" /etc/

Sample outputs:

# ms-wins 192.168.1.50
# ms-wins 192.168.1.51
addresses1=192.168.1.5;24;192.168.1.2;

How to use grep to search words only

When you search for boo, grep will match fooboo, boo123, barfoo35 and more.
You can force the grep command to select only those lines containing matches
that form whole words i.e. match only boo word:

$ grep -w "boo" file

5 of 12
How to use grep to search 2 different words
Use the egrep command as follows:

$ egrep -w 'word1|word2' /path/to/file

Ignore case
We can force grep to ignore case distinctions in patterns and data. For example,
when I search for ‘bar’, match ‘BAR’, ‘Bar’, ‘BaR’ and so on:

$ grep -i 'bar' /path/to/file

In this example, I am going to include all subdirectories in a search:

$ grep -r -i 'main' ~/projects/

How can I count line when words has been

matched
The grep can report the number of times that the pattern has been matched for
each file using -c (count) option:

$ grep -c 'word' /path/to/file

Pass the -n option to precede each line of output with the number of the line in
the text file from which it was obtained:

6 of 12
 $ grep -n 'root' /etc/passwd

1:root:x:0:0:root:/root:/bin/bash
1042:rootdoor:x:0:0:rootdoor:/home/rootdoor:/bin/csh
3319:initrootapp:x:0:0:initrootapp:/home/initroot:/bin/ksh

Force grep invert match

You can use -v option to print inverts the match; that is, it matches only those
lines that do not contain the given word. For example print all line that do not
contain the word bar:

$ grep -v bar /path/to/file

$ grep -v '^root' /etc/passwd

Display lines before and after the match

Want to see the lines before your matches? Try passing the -B to the grep:

grep -B NUM "word" file

grep -B 3 "foo" file1

Similarly, display the lines after your matches by passing the -A to the grep:

grep -A NUM "string" /pth/to/file

grep -A 4 "dropped" /var/log/ufw.log

7 of 12
We can combine those two options to get most meaningful outputs:

grep -C 4 -B 5 -A 6 --color 'error-code' /var/log/httpd

/access_log

Here is a sample shell script that fetches the Linux kernel download urls:

.......
...
_out="/tmp/out.$$"
curl -s https://www.kernel.org/ > "$_out"
#######################
## grep -A used here ##
#######################
url="$(grep -A 2 '<td id="latest_button">' ${_out} |
grep -Eo '(http|https)://[^/"]+.*xz')"
gpgurl="${url/tar.xz/tar.sign}"
notify-send "A new kernel version ($remote) has been released."
echo "* Downloading the Linux kernel (new version) ..."
wget -qc "$url" -O "${dldir}/${file}"
wget -qc "$gpgurl" -O "${dldir}/${gpgurl##*/}"
.....
..

UNIX / Linux pipes

grep command often used with shell pipes. In this example, show the name of
the hard disk devices:

# dmesg | egrep '(s|h)d[a-z]'

Display cpu model name:

# cat /proc/cpuinfo | grep -i 'Model'

However, above command can be also used as follows without shell pipe:

8 of 12
 # grep -i 'Model' /proc/cpuinfo

model : 30
model name : Intel(R) Core(TM) i7 CPU Q 820 @ 1.73GHz
model : 30
model name : Intel(R) Core(TM) i7 CPU Q 820 @ 1.73GHz

One of my favorite usage of grep or egrep command to filter the output of the
yum command/dpkg command/apt command/apt-get command:

dpkg --list | grep linux-image

yum search php | grep gd
apt search maria | egrep 'server|client'

9 of 12
 Linux grep commands explained with shell pipes examples

How do I list just the names of matching files?

Use the -l option to list file name whose contents mention main():

$ grep -l 'main' *.c

OR

$ grep -Rl 'main' /path/to/project/dir/

Colors option
Finally, we can force grep to display output in colors, enter:

$ grep --color vivek /etc/passwd

10 of 12
 Grep command in action

In conclusion, the --color option increase readability. For example, the

GREP_COLOR environment variable and the grep --color=always can be used as
follows:

GREP_COLOR='1;35' grep --color=always 'vivek' /etc/passwd

GREP_COLOR='1;32' grep --color=always 'vivek' /etc/passwd
GREP_COLOR='1;37' grep --color=always 'root' /etc/passwd
GREP_COLOR='1;36' grep --color=always nobody /etc/passwd

In addition, to default red color now we can define colors using GREP_COLOR
shell variable. The differnt color helps us massivly with visual grepping.

Conclusion

11 of 12
The grep command is a very versatile and many new Linux or Unix users find it
complicated. Hence, I suggest you read the grep man page too. Let us
summarize most import options:

Linux grep command options Description

Ignore case distinctions on Linux and
-i
Unix
Force PATTERN to match only whole
-w
words
-v Select non-matching lines
-n Print line number with output lines
Suppress the Unix file name prefix on
-h
output
-r Search directories recursivly on Linux
-R Just like -r but follow all symlinks
Print only names of FILEs with selected
-l
lines
Print only a count of selected lines per
-c
FILE
--color Display matched pattern in colors

12 of 12