CODEWITHSUNDEEP
regexbashsed
rualonor
rualonor

Reputation: 31

Sed wildcards- replace in the middle of certain characters

I have a line like this

"RU_28" "CDM_279" "CDI_45"
"RU_567" "CDM_528" "CDI_10000"

I want to obtain the result below

"RU_28" "CDM_Unusued" "CDI_45"
"RU_567" "CDM_Unusued" "CDI_10000"

Do this for all the lines in the file

I'm using this commands: 

sed 's/\"CDM_\w*\"/\"Unusued\"/g' File1.txt > File2.txt

It doesn't seem to works.

Thanks in advance!!!

Upvotes: 3

Views: 5133

Answers (2)

Dan Fego
Dan Fego

Reputation: 14014

You're not actually leaving "CDM_" on the right side. Your substitution says "Replace CDM_ and any number of words with Unusued." The common way to do what you actually want to do is to use parentheses around the section on the left you want to keep, and then use backreferences to indicate where they go on the right. In this case, you just need a single backreference, indicated by \1:

sed 's/"\(CDM_\)\w*"/"\1Unusued"/g' File1.txt > File2.txt

Note the backslashes before the parentheses on the left side; these can be omitted if using sed with -r (I think), for extended regexps, but as-is, they're necessary so sed knows they're not literal.

Edit: I've updated the command in response to the accurate comment by Birei, noting that the extraneous escapes for the double-quotes. (Note that the ones for the parentheses are still necessary).

Upvotes: 1

anubhava
anubhava

Reputation: 785541

You can use:

sed -i.bak 's/"\(CDM_\)[^"]*"/"\1Unused"/' file1.txt

Upvotes: 2

Related Questions