[SOLVED] Bash Script remove line breaks and carriage returns

buee · 06-06-2012, 04:27 PM

I have the need to turn this:

Code:

 1   server=hotspot1 name="tsmith" password="password" profile=package5 
     uptime=22m45s bytes-in=2238756 bytes-out=14770276 packets-in=14837 
 2   server=hotspot1 name="jbob" password="password" profile=package5 
     uptime=3m3s bytes-in=92196 bytes-out=952379 packets-in=849

into this (a few hundred times):

Code:

 1   server=hotspot1 name="tsmith" password="password" profile=package5 uptime=22m45s bytes-in=2238756 bytes-out=14770276 packets-in=14837 
 2   server=hotspot1 name="jbob" password="password" profile=package5 uptime=3m3s bytes-in=92196 bytes-out=952379 packets-in=849

Unfortunately this is an excerpt from a file generated by a router and I do not have any control over how the file is formatted.

I used this:

Code:

sed -e :a -e '$!N;s/\r\n     //;ta'

and that produces what I want, but for group 1 only, it doesn't repeat for group 2, group 3, etc. I also tried running that command in a while read $line loop:

Code:

     uptime=22m45s bytes-in=2238756 bytes-out=14770276 packets-in=14837 
 2   server=hotspot1 name="jbob" password="password" profile=package5 
     uptime=3m3s bytes-in=92196 bytes-out=952379 packets-in=849

And it appears to delete the first line of the text, then no affect after that.

Anyone know how to accomplish this?

sycamorex · 06-06-2012, 06:01 PM

Is that what you're after?

Code:

sed '/server/ N;s/\n     //g' file

Please note there are some spaces after \n
I don't know if they are in the original output. They do exist in what I copied from your post.

buee · 06-07-2012, 08:09 AM

Quote:

Originally Posted by sycamorex

Is that what you're after?

Code:

sed '/server/ N;s/\n     //g' file

Please note there are some spaces after \n
I don't know if they are in the original output. They do exist in what I copied from your post.

Not quite. That ended up deleting up to:

Code:

uptime=22m45s bytes-in=2238756 bytes-out=14770276 packets-in=14837

On each iteration.

grail · 06-07-2012, 08:47 AM

How about:

Code:

awk '{ORS=($1~/^[0-9]/)?" ":"\n"}$1=$1' file

buee · 06-07-2012, 09:38 AM

This is probably one of the worst possible ways to do it, but I finally got it figured out and formatted the way I need it:

Code:

while [ $i -le $lines ] ; do
        cat temp$i | head -2 | sed -e :a -e '$!N;s/\r\n     //;ta' >> hsoutput.txt
        sed '1,2d' temp$i > temp$(( $i + 1 ))
        rm -rf temp$i
        i=$(( $i + 1 ))
done

Then I just do an rm -rf temp* at the end of the script to clean up the extra garbage.