Hi there!

This sounds like such a simple task, yet I haven't succeeded so far... I have some standard, not-so-long linux text file, say, readme.txt, and I want to delete all new lines (\n). How can this be so difficult?

I tried
but this doesn't work. I guess, that's because sed reads only one line per go? So I could use
which would join adjacent lines all right. But, of course, to replace all line breaks, I would have to do this more often
or even put a loop around it. But this seems so difficult! Ain't there any better way?

(actually, what I want to do is to replace the standard "You have Mail in /var/spool/..." message by "You have Mail from Tom, Andi, Sue". So I'm looking for a fast one-liner )

Cheers,
Bkeeper

For not-so-long linux text file, you can do:

VAR=`cat readme.txt` && echo $VAR

you get one long concatenated line where the \n are replaced by a blank character.

to remove this space, you have to mark the ends of line, like that:

VAR=`sed 's/$/$$/' readme.txt` && echo $VAR ' '|sed 's/\$\$ //'g

Yes, that works!
Thank you very much.
Cheers,
BKeeper

here's the sed command you were looking for:

sed ':a;N;$!ba;s/\n//g' <inputfile>

The ':a' is a marker.
The '$!ba' causes sed to branch to the ':a' marker for all lines except the last line. This creates one huge line.
Finally, all \n's are removed.

This took me a while to find, but it's one of those things you'll want to keep around.

Pablo

Hi Pablo,
great!
Thanks for you post. It'll take me a bit to understand the expression, but, well, you know, with sed it's always worth the effort...
Cheers,
BKeeper