Shell Programming and Scripting

Hi,

i want to replace some lines in a file "Source.txt" by taking search pattern input from a different file named "abc".

Search pattern input file have following lines:

vi abc
2637918
2707239
2709124
2709148
2709159
2709177
...........

i am running following code:

for i in `cat abc` ;do perl -pi.bk4 -e 's/^.*$i.*//g' Source.txt;done

but it's not working.......
pls help.

while read line
do
sed -i "s/^.*$line.*//g" source.txt //or// sed -i 's/^.*${line}.*//g'
done < abc

or try

You want to replace them with empty lines, or delete them? Delete is easy:

Code:

fgrep -vf abc Source.txt

For replacing with nothing, perhaps the following:

Code:

sed -e 's%.*%s/.*&.*//%' abc | sed -f - Source.txt

Maybe your sed can't handle standard input for the -f option; if so, report back here and we'll think about workarounds.

Repeated looping over all the input patterns seems awfully inelegant and inefficient.

Hi Vidya,
ur sed commnd is replacing the input seach pattern line with nothing but.... size of output file is getting 3 times more than the orignal file.... means each time after replacing the matched line with nothing the command is adding the entire file into the output file again....

pl help anybody.......hw to overcome this problem...

Code:

while read line
do
perl -i -ne 's/'$line'//g; print; ' filename
done < abc

Dennis...it's also not working.......pls specify clearly.

suppose my output file is OUTFILE .....

pls tell me the exact code..

Its working well!!!!
Input:

Code :

Output: