Grep is a command line tool that allows you to find a string in a file or stream.
How To Use grep
In the simplest case, grep can simply be invoked like this :
grep 'STRING' filename
This is OK but it does not show the true power of grep. First this only looks at one file. A cool example of using grep with multiple file would be to find all files in a directory that contains the name of a person. This can be easily accomplished using a grep in the following way :
grep 'Nicolas Kassis' *
Notice the use of single quotes; the quotes are not essential, but in this example they are required because the name contains a space. Double quotes could also have been used in this example.
Few examples (of grep command):
grep –v abc file1
(don’t show line that contain abc but all)
grep -n tiger file1
(return line that contain tiger with its line number)
grep –c abc file1
(display count of line only)
grep -l wali *
[only returns the name of file that have at least one line containing the patter (wali)]
<root@wtuto: ~> # grep -r abc *
(perform a recursive search of files, starting with the named directory)
(Note: there are there file f1, f2, f3 in /root/wali (directory) and f2 doesn’t contain the word abc. )
<root@wtuto: wali> # grep -r rose *
[perform case sensitive search (in the current directory)]
Extracting text by column:
cut -f3 -d: /etc/passwd (display third colon(:)-delimited field)
cut -c5 /etc/passwd (display 5th character)
cut -c1-5 /etc/passwd (display first 5 characters)
-d specify the column delimiter (default is TAB)
-f specify column to print
-c cut by character
Gathering text statistics:
wc file1
(displays no. of lines, words and character in file1)
wc *
(displays no. of lines, words and character of every files in the current directory)
|
-l |
only for line count |
|
-w |
only for word count |
|
-c |
only for byte count |
|
-m |
only for character count (1 character = 1 byte) |
Sorting text:
grep bash /etc/passwd | sort
sort -t : -k3 -n /etc/passwd
(sort the UIDs in ascending order)
sort -t : -k3 -n /etc/passwd | cut -f3 -d:
(shows only UIDs in ascending order)
|
-r |
performs a reverse (descending) sort |
|
-n |
performs a numeric sort |
|
-f |
ignores (folds) case of characters in strings |
|
-u |
(unique) removes duplicate lines in output |
|
-t : |
uses : as a filed separator |
|
-k 3 |
third column by : delimited field |
Eliminating duplicate lines:
cat>file
wali
salman
obama
wali
wali
wali
ajay
sameer
uniq file
(uniq without argument removes duplicate adjacent lines)
wali
salman
obama
wali
ajay
sameer
-u to output only the lines that are truly unique, only occurring once in the input.
uniq -u file
(display following)
wali
salman
obama
ajay
sameer
-d to output only print one copy of the lines that are repeated in the input
uniq -d file
(show following)
wali
-c
each line will be prepended with a
number indicating how many times it appears in the input.
uniq -c file
(display following)
1 wali
1 salman
1 obama
3 wali
1 ajay
1 sameer
Comparing files:
diff and sdiff compares two files for differences.
cat>file
wali
salman
obama
wali
wali
wali
ajay
sameer
cp file file1
vim file1
wali
salman
president
wali
wali
wali
ajay
sameer
diff file file1
diff -u file file1
(line that begin with + exist in file but not in file1,
line that begin with - exist in file1 but not in file)
Spell checking with aspell:
aspell check f1
Tools for manipulating text:
cat f1 | tr 'a-z' 'A-Z'
(convert lower case into uppper)
sed 's/she/he/' f1 > nf1
(replace she with he in f1 and send
its result to nf1, but only one change per line, and is case sensitive)
sed 's/she/he/g' f1 > nf1
(g stand for global, which replace all she with he in f1 and send its result to nf1, and is case sensitive)
sed 's/she/he/gi' f1 > nf1
(i ignore case sensitive, rest is same)
sed '2,5s/she/he/gi' f1 > nf1
(replacement occure only from 2nd to 5th line)
sed '/abc/,/xyz/s/she/he/gi' f1 > nf1
(replacement occure only from the line where abc occure to the line where xyz occure)
