Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems. It only takes a minute to sign up.
Sign up to join this community
I am trying with the following sed command:
sed 's/shared:core:([0-9]{1,4}(\.[0-9a-z]{1,6}){1,5})/shared:core:1.1.1/' test.txt
to replace the content of test.txt
shared:core:0.0.2
I expect it to become
shared:core:1.1.1
But nothing is happening. I think I need a little hint on how that works with sed
You are using (foo) which is mean to capture the string foo. However, by default, sed will read that as a literal ( followed by foo and another ) and since you don't have any ( or ) in your input, that will never match.
For sed to use capturing parentheses, you need t escape them (\(foo\)) or you need to use -E. You don't actually need this at all here though because you are never referring to the thing you were trying to capture. All you need is:
$ sed 's/shared:core:.*/shared:core:1.1.1/' file
shared:core:1.1.1
Or even:
$ sed -E 's/(shared:core:).*/\1:1.1.1/' file
shared:core::1.1.1
Or even simpler:
$ sed -E 's/(.*):.*/\1:1.1.1/' file
shared:core:1.1.1
Now, if your input is more complicated and what you are trying to match is 1 to 5 repetitions of 1-4 numbers followed by one to six repetitions of a . followed by one to 6 alphanumeric characters (which is what the regex you used does), you can use your version, just add -E (assuming your sed implementation supports it):
$ sed -E 's/shared:core:([0-9]{1,4}(\.[0-9a-z]{1,6}){1,5})/shared:core:1.1.1/' file
shared:core:1.1.1
Or, if you can't use -E:
$ sed 's/shared:core:\([0-9]\{1,4\}\(\.[0-9a-z]\{1,6\}\)\{1,5\}\)/shared:core:1.1.1/' file
shared:core:1.1.1
Although I fear this might not be 100% portable, not sure.
Finally, remember that sed doesn't change the input file unless you give it the -i option. So if you are expecting the original file to change, use sed -i .... file.
By default, sed uses basic regular expressions. Add the -E option for extended regular expressions. Thus:
sed -E 's/shared:core:([0-9]{1,4}(\.[0-9a-z]{1,6}){1,5})/shared:core:1.1.1/' test.txt
Alternatively, convert to a basic regular expression. Thus:
sed 's/shared:core:\([0-9]\{1,4\}\(\.[0-9a-z]\{1,6\}\)\{1,5\}\)/shared:core:1.1.1/' test.txt
Additionally, if you actually want the contents of the file to change, then you need to add the -i option, to cause an "in place" change. Thus:
sed -E 's/shared:core:([0-9]{1,4}(\.[0-9a-z]{1,6}){1,5})/shared:core:1.1.1/' -i test.txt
sed -i-E ... will save the old file like test.txt-E, and will not use the extended regular expressions. What you want is sed -i -E ... Notice the space between -i and -E.
-i option is a more recent addition that allows a file to be changed "inplace". (-E is also more recent.)
.followed by one to 6 alphanumeric characters". But this doesn't make much sense with the input example you have given.diffing it with a reference version).sed 's/0\.0\.2/1.1.1/? If so, why do you have such a complicated regex? Are there any more lines in the file that you need to avoid?