JIYIK CN >

In this article, let's review some interesting workarounds using the "s" substitution command in sed with a few real-world examples.

1. sed replaces the delimiter

As we discussed in our previous article, we can use different delimiters in sed substitution command, @ % | ; :e.g.

Let us first create the path.txt file which will be used in all the examples mentioned below.

path.txt

/usr/kbos/bin:/usr/local/bin:/usr/jbin:/usr/bin:/usr/sas/bin
/usr/local/sbin:/sbin:/bin/:/usr/sbin:/usr/bin:/opt/omni/bin:
/opt/omni/lbin:/opt/omni/sbin:/root/bin

Example 1 – sed @ delimiter: Replace /opt/omni/lbin with /opt/tools/bin

When we replace /a path name that contains , we can use @as a delimiter instead of /. In the following sed example, in the last line of the input file, **/opt/omni/lbin** has been changed to /opt/tools/bin .

$ sed 's@/opt/omni/lbin@/opt/tools/bin@g' path.txt

Example 2 – sed / delimiter: Replace /opt/omni/lbin with /opt/tools/bin

When we should use in pathname related replacement /, we have to escape in replacement data /as shown below. In this sed example, the delimiter is escaped /in both REGEXP and REPLACEMENT parts.

$ sed 's/\/opt\/omni\/lbin/\/opt\/tools\/bin/g' path.txt

2. sed &gets the matching string

The exact part of the input line that the regular expression matched is &represented by and can then be used in the replacement part.

Example 1 – sed & Usage: Replace /usr/bin/ with /usr/bin/local

In the following example, in the replacement part, /usr/bin& matching the pattern will be replaced and added to /local . So in the output all occurrences of /usr/bin will be replaced with /usr/bin/local .

$ sed 's@/usr/bin@&/local@g' path.txt

Example 2 – sed & usage: matching the entire line

&Replace any matches with the given REGEXP .

$ sed 's@^.*$@<<<&>>>@g' path.txt

In the above example, the regular expression ^.*$matches the entire line. The replacement part is written in the entire line using and <<<&>>>at the beginning and end of the line respectively .<<<>>>

3. Grouping and backreferences in Sed

Groups can be used in sed just like normal regular expressions. Groups \(are turned on with , and \)off with . Groups can be combined with backreferences.

A backreference is a reuse of a portion of a regular expression selected through grouping. Backreferences in sed can be used in both the regular expression and the replacement part of the substitute command.

Example 1: Get only the first path in each line

In the following example, the path available before \(\/[^:]*\)the first colon is matched . The first matched group is replaced.:\1

$ sed 's/\(\/[^:]*\).*/\1/g' path.txt

Example 2: Multiple Groups

In the file path.txt, change the order of the fields in the last line of the file.

$ sed '$s@\([^:]*\):\([^:]*\):\([^:]*\)@\3:\2:\1@g' path.txt

In the above command, $you specify that replacement should be done only on the last line. The output shows that the order of the path values ​​in the last line is reversed.

Example 3: Get a list of usernames in the /etc/passwd file

This sed example displays only the first field in the /etc/passwd file.

$ sed 's/\([^:]*\).*/\1/' /etc/passwd

Example 4: Add brackets around the first character of each word

This sed example adds brackets before the first character of each word.

$ echo "Welcome To The Jiyik.com" | sed 's/\(\b[A-Z]\)/\(\1\)/g'

Example 5: commify a simple number.

Let us create a file called numbers which contains a list of numbers. The following sed command example is used to group the numbers up to thousands.

numbers

1234
12121
3434
123

Look at the following command

$ sed 's/\(^\|[^0-9.]\)\([0-9]\+\)\([0-9]\{3\}\)/\1\2,\3/g' numbers