CODEWITHSUNDEEP
regexunixsed
Roy
Roy

Reputation: 69

Having trouble to understand a portion of regex

This is the text:

We will now present the Linux ls command
   ... here description of ls

We will now present the Linux cd command
   ... here description of cd
   ... more description
Done

This following sed replacement regex is applied on the text

sed 's/.*Linux \(.*\) .*/\1:/' ex2.txt

Which gives the following output

ls:
   ... here description of ls

cd:
   ... here description of cd
   ... more description
Done

Can someone please walk me through on how it worked?

Upvotes: 0

Views: 104

Answers (1)

Sebastian Simon
Sebastian Simon

Reputation: 19485

.*Linux : Anything before the word “Linux” and that word itself followed by a space. (We will now present the Linux )

\(.*\): Then something captured with parentheses. (ls or cd)

 .*: Then another space and anything else. ( command)

Replaced by:

\1:: Whatever the first capturing group has captured, followed by a colon (ls: or cd:)

In this case, . matches any character except line breaks.

\(.*\) matches only that one word because there has to be a space before “Linux” and there has to be a space after that word: the space between the command name and the word “command”.

The parentheses don’t have to be escaped with a backslash if extended regex is used (sed -r…).

So you could also write:

sed -r 's/.*Linux (.*) .*/\1:/' ex2.txt

Upvotes: 3

Related Questions