How to replace part of a regex with sed?

Question

In a file test.txt I have the following text line

COPYRIGHT (c) 2020 alex4200

in which I want to use sed to find that line and to replace the year 2020 by the year 2021. I have tried the following expression:

sed -i -E  "s/COPYRIGHT .*(\d\{4\}) alex4200/2021/" test.txt

but it did not change the text in the file test.txt. What am I missing?

You are searching for a literal {, which is not present in your string. Also, \d has no special meaning in a [POSIX regular ](regular-expressions.info/posix.html), not even in extended POSIX. — user1934428
– user1934428, Commented May 26, 2021 at 6:27

RavinderSingh13 · Accepted Answer · 2021-05-26 06:42:27Z

2

First thing first \d is not supported in sed you need to use [0-9]. Then your expression is also not correct(even you fix \d it will not work with your tried code), so I have corrected it as per your shown samples. With your shown samples, please try following.

sed -E 's/^([^ ]* +[^ ]*) +([^\s]*) +(.*)$/\1 2021 \3/' Input_file

Once you are happy with results(which will be shown on terminal) you can use -i option to do inplace save in above code.

Explanation: Simple explanation would be, using sed's backreference capability while substitution. In first part of substitution part, making 3 back references as per shown samples, where 2nd back reference will have 2020 value in it, while doing substitution putting 2021 value there.

Explanation of regex:

^([^ ]* +[^ ]*)  ##From starting of value creating 1st back reference which will match values just before 2nd occurrence of space.
 +               ##Matching 1 or more occurrences of spaces here.
([^\s]*)         ##Creating 2nd capturing group which will match everything till space comes, to catch 2020 basically.
 +               ##Matching 1 or more occurrences of spaces here.
(.*)$            ##Matching everything till last here and creating 3rd back reference here.

Fixing OP's attempts:

sed -E  's/^(COPYRIGHT .*\(c\))\s+([^ ]*)\s+(.*)/\1 2021 \3/' Input_file

OR to match digits with shown samples, following could be used:

sed -E 's/^(COPYRIGHT\s+\(.*\))\s+([0-9]{4})(.*)$/\1 2021 \3/' Input_file

edited May 26, 2021 at 6:42

answered May 26, 2021 at 6:22

RavinderSingh13

135k14 gold badges61 silver badges100 bronze badges

Sign up to request clarification or add additional context in comments.

10 Comments

Alex Over a year ago

I want to change the content of the file test.txt. Maybe you misunderstood my question?

RavinderSingh13 Over a year ago

@Alex, I have added a note about -i once you are Good with my code, which will print on terminal you can use -i option to do inplace save into Input_file.

Alex Over a year ago

The text I want to search and replace is in a file, not in an echo statement. So I have to do something like echo cat test.txt` | sed -i -E 's/^([^ ]* +[^ ]*) +([^\s]*) +(.*)$/\1 2021 \3/'` That does not work, I get an error 'sed: No input file`

RavinderSingh13 Over a year ago

@Alex, sure, have edited my code now to take input from a Input_file.

RavinderSingh13 Over a year ago

@Alex, sure then sed -E 's/^(COPYRIGHT\s+$.*$)\s+([0-9]{4})(.*)$/\1 2021 \3/' file will work. Once you confirm this I will add it in my answer.

|

Collectives™ on Stack Overflow

How to replace part of a regex with sed?

1 Answer 1

10 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

10 Comments

Your Answer

Sign up or log in

Post as a guest

Related