Shell Programming and Scripting

Hello All,
I have a bunch of files that have the following format, where the title is INPUT.txt and contains the following text:

Code:

INPUT-FILLER1 204
INPUT-FILLER2
FILLER6-INPUT 5
FILLER-INPUT

I want to go through the directory and remove the keyword INPUT. For example, my output would be

Code:

FILLER1 204
FILLER2
FILLER6
FILLER

I have multiple files where the keyword is the title of the .txt file, so I figured bash would be the way to go. I was thinking I would need two sed statements. One to remove the INPUT and the other to remove "-"

Any help would be appreciated.

If replacing all strings of the form "INPUT-" or "-INPUT" regardless of where they appear in the file will work for you, then this will do the trick:

Code:

sed 's/INPUT-//; s/-INPUT//' input-file >output-file

If the INPUT- strings appear only at the beginning of the line, then make the first substitute command this: s/^INPUT-//.


Test this carefully -- Not being sure what else your file contains, I'm not 100% sure that this won't replace something meaningful.

thanks. I had to adjust your code slightly (< input file).

I modified it quite a bit to fit into a bash script, but i'm having issues. Any ideas?

Try

Code:

for i in directory/*.txt
do
    var=$(echo $i |cut -d . -f 1) # no '.txt' after $i
    sed "s/$var-//; s/-$var//" < "$var".txt  > "$var".tmp
done

yazu/agama, thanks for your continued help.

I tried your suggestion and got a repeated error: "sed: -e expression #1, char 22: unknown option to `s'"

Debug your script to see what you get:

Code:

for i in directory/*.txt
do
    var=$(echo $i |cut -d . -f 1)
    echo "$var"
    echo sed "s/$var-//; s/-$var//"
done