My script takes a file name in the form R#TYPE.TXT (# is a number and TYPE is two or three characters). I want my script to give me TYPE. What should I do to get it? Guess I need to use awk and sed.
I'm using /bin/sh (which is a requirement)
Add a comment
|
3 Answers
you can use awk
$ echo R1CcC.TXT | awk '{sub(/.*[0-9]/,"");sub(".TXT","")}{print}'
CcC
or
$ echo R1CcC.TXT | awk '{gsub(/.*[0-9]|\.TXT$/,"");print}'
CcC
and if sed is really what you want
$ echo R9XXX.TXT | sed 's/R[0-9]\(.*\)\.TXT/\1/'
XXX
Comments
I think this is what you are looking for.
$ echo R3cf.txt | sed "s/.[0-9]\(.*\)\..*/\1/"
cf
If txt is always upper case and the filename always starts with R you could do something like.
$ echo R3cf.txt | sed "s/R[0-9]\(.*\)\.TXT/\1/"
3 Comments
magol
If I use you example, I get sed: -e expression #1, char 20: invalid reference \1 on `s' command's RHS I rewrite it to: sed "s/R1(.*)\.TXT/\1/"
William Clemens
I believe that if you write it RegEx like that your file name is going to have to start with "R1". Whats your "uname -a"?
magol
upps, took the wrong code. Should be sed "s/R[0-9](.*)\.TXT/\1/"
You can use just the shell (depending what shell your bin/sh is:
f=R9ABC.TXT
f="${f%.TXT}" # remove the extension
type="${f#R[0-9]}" # remove the first bit
echo "$type" # ==> ABC
1 Comment
magol
On the old unix system I code for I get "bad substitution" when I type f="${f%.TXT}"