1

I want to add $x/ from data formatted as $x-xyz (xyz is a mix of text and number to be ignored).

Example data

AB-10C
CD-02E

Output

AB/AB-10C
CD/CD-02E

How do I describe the variable as text before the hyphen? My idea is:

define $x = # don't know how to do that
## and then somehow use sed to prepend
sed -i s/$/$x/ file
# I don't know if sed can handle variables though - the above one liner is from Googling

2 Answers 2

5

Here's a quick and dirty awk version:

awk -F- '{print $1"/"$0}' input_file > output_file

What it does is use - as a field separator, and prints the first column (i.e. everything before the first -), then a /, then the whole original line.

A way of doing the same thing with sed would be:

sed -e 's;^\([^-]*\)\(.*\);\1/\1\2;' input_file > output_file

(but that's hardly readable.)

If you want to do it in plain bash, you can use string manipulations:

$ foo=AB-10C
$ prefix=${foo%%-*}
$ echo ${prefix}/${foo}
AB/AB-10C

Use that in a while read loop or similar if the data is coming from a file.

2
  • Though it was unintentional, I simplified the example a bit, and the results with the real data were interesting. The actual data was like 211,AB-45A,1, and it got manipulated to 211,AB/211,AB-45A,1 with both awk and sed. Couldn't try pure bash since I'm not quite familiar with while read and didn't bother researching. Anyway, I separated the relevant data and it worked as expected - thank you!
    – Oxwivi
    Jan 29, 2012 at 14:19
  • For the while loop, just type "[bash] while read" in the search box at the top of this site. You'll get plenty of examples :)
    – Mat
    Jan 29, 2012 at 14:21
0

This might work for you:

echo -e "123,AB-10C,4\n567,CD-02E,8" |
sed 's|\([^,]*\)-.*|\1/&|'
123,AB/AB-10C,4
567,CD/CD-02E,8

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .