CODEWITHSUNDEEP
fileunixsedfindline
Luke
Luke

Reputation: 7089

Delete all reoccurrences of first line in a file using sed

I've got a bunch of files with multiple rows containing copies of the header. Is there a way using sed to delete all occurrences except for the first line. I was thinking something like:

sed -i '/textOnlyInHeader/d' file.txt

Except this will delete the header as well. As a bonus, is there a way to do this recursively for all files in a bunch of subdirectories?

Upvotes: 1

Views: 86

Answers (4)

potong
potong

Reputation: 58401

This might work for you (GNU sed):

sed '1h;1b;G;/^\(.*\)\n\1$/!P;d' file

Upvotes: 2

NeronLeVelu
NeronLeVelu

Reputation: 10039

sed -i '1 !{
   /textOnlyInHeader/ d
   }' file.txt

avoid first line and use your sed on other line

for the recursive, you could passe to sed a list of file (in place of just file.txt). So prepare the list before with a shell function (find, ls, ..., loop) an pass it to the sed as argument

Upvotes: 1

mpez0
mpez0

Reputation: 2883

I know there's already an accepted answer using gawk, but using sed:

sed -i -e '2,$s/textOnlyInHeader/DELETELINE/' -e '/DELETELINE/d' file.txt

for the recursive answer, I concur with Steve that a loop with find is the way to go.

Upvotes: -1

Steve
Steve

Reputation: 54392

I think gawk would be best for this. Try:

gawk -i inplace 'NR==1 { r = $0; print } r == $0 { next }1' file.txt

For all files in a single directory, change NR to FNR and run:

gawk -i inplace '...' *.txt

For all files in many subdirectories, you can use a for loop:

for i in $(find /path/to/files -type f -name '*.txt'); do ... ; done

If you're using an old or non-GNU AWK, you will need to write to a temp file first:

awk '...' file.txt > file.tmp && mv file.tmp file.txt

Upvotes: 3

Related Questions