Hi! I''ve got a file with a structure like this: 1:Eeee Rrr:Ttt Yyyy: 2:Rrrr Ffff:Ggggg hhhh jjj: etc. What should I do in sed to: - find the line with a specific number in the beginning - write a string at the end of this line (after the last ":")? I beg for help.... -------------- "We cannot all be masters." - William Shakespeare

Not sed but works anyway:

to find line '5' and write 'foo' at the end (and save the original unmodified file as .bak):


perl -pi.bak -e 's/^5(.*):$/5$1:foo/' file.txt

HTH

[edited by - grazer on May 26, 2004 1:41:39 PM]

Here's some sed that works on Windows XP.
First, if you want just the one line with the number that you specify (not the rest of the file) try this:

sed -n "s/^2\(:.*:\)$/2\1sed is fun/p" "in.txt"  

Prints

2:Rrrr Ffff:Ggggg hhhh jjj:sed is fun  

If you want the whole thing but with the one line modified, try

sed "s/^2\(:.*:\)$/2\1sed is fun/" "in.txt"  

In the first example, the -n suppresses printing, and the p after the last / prints lines with a match. I use the GNU sed manual (at http://www.gnu.org/software/sed/manual/html_mono/sed.html) for a reference.

Enjoy.

[edited by - Metaphorically on May 26, 2004 7:25:17 PM]

Thanks, Metaphorically.
However I''ve already solved my problem, though a bit way around. I simplified it, but maintaining its job.
The funny thing is, that in the system (a UNIX) I''m doing it on there''s an undocumented parameter for sed: -i. This parameter has sed modify a file, without printing to the screen. I don''t have to use any ''w'' option with it. Strange, but actually nice.
Still, thanks for your post, since I always love to see some regular expressions working. I have a test (bash & awk) next week and I take any chance of learning anything.

--------------
"We cannot all be masters." - William Shakespeare