Unix command help: find string pattern and delete line

cronjob78

Need to delete a line in a large number of files that contains a certain pattern. The file line that the pattern appears varies from file to file. The line that will contain the pattern will also vary slightly. IM sure you can probably do this with sed but I dont know how to pipe the line number from the find command into sed to delete the line.

In pseudocode:

If line of file contains partial string 'abcdef'
then delete that line.

Ideally Id like to be able to run this from the command line.


Any help appreciated.

djmarkus

cronjob78 wrote: »

Need to delete a line in a large number of files that contains a certain pattern. The file line that the pattern appears varies from file to file. The line that will contain the pattern will also vary slightly. IM sure you can probably do this with sed but I dont know how to pipe the line number from the find command into sed to delete the line.

In pseudocode:

If line of file contains partial string 'abcdef'
then delete that line.

Ideally Id like to be able to run this from the command line.


Any help appreciated.

grep -v abcdef file

Catch_22

for file in `find . -type f`
do
grep -v abcdef $file >/tmp/$$
mv /tmp/$$ $file
done

Orion

and output to another file:

grep -v abcdef file > newfile

or script it to

grep -v "$1" "$2" > outputfile
rm "$1"
mv outputfile "$1"

[edit]or catch_22's solution for all files in a dir.

E39MSport

try: -
sed -e '/abcdef/d' filename > newfile
mv newfile filename

if you have a bunch of files, you could write a small shell like
fileList="`ls ${fileDir}/*`"
for yourFile in $fileList ; do
sed -e '/abcdef/d' $yourFile > $newFile
mv $newFile $yourFile
done

google is your best friend for stuff like this

Skrynesaver

perl -ei 'while(<>){print $_ unless (/REGEX/)};' FILENAMES