Does any one know how to apply sed line by line and not directly on a file text :
for i in $(cat server); do
Exclude_List="AF|PO"
echo $i | egrep $LBB3B
if [ $? -eq 0 ]; then
**do my sed on the $i line only then continue on next line**
fi
done
sed already works line by line:
sed -E '/AF|PO/{ ...sed expression to apply when matching... }' server
For example, only print the lines matching that regular expression:
sed -nE '/AF|PO/p' server
The -E flag to sed makes the utility interpret the regular expression as an extended regular expression, and the -n turns off the implicit outputting of every input line.
For doing multiple things to lines matching the expression, enclose these in { ... }. For example, print each matching line twice (and don't print non-matching lines):
sed -nE '/AF|OP/{p;p;}' server
sed has line addressing capabilities, if that helps you:
$ seq 5 > input
$ sed '3s/.*/jeff/' input
1
2
jeff
4
5
$ sed '3,5s/.*/new/' input
1
2
new
new
new
Instead of using a shell loop to process a text file, consider a tool that more naturally processes text, such as awk:
$ cat input
AF bar
PO baz
other stuff
more other stuff
$ awk '/(AF)|PO/ { print }' input
AF bar
PO baz
$ awk '/stuff/ && !/more/ { print }' input
other stuff
That's not what sed is for. You can use the native abilities of the shell do everything "line-by-line".
sedwith here-stringsed .... <<<"$i"as well as piping to itecho "$i" | sed ..... – αғsнιη Mar 21 '18 at 16:21sedfor something, you need to figure out whether it's the correct tool for the job.sedis a *stream* editor, that's stream editor. It's not designed for doing what you want. Don't get me wrong; you could wire something up Rube Goldberg style, but don't. Just pick the correct tool for the job. See my answer below. – user1404316 Mar 21 '18 at 16:22