ksh help: to extract 2 words from every line in a file

[Hanry Zhou]

I have a lot of lines in a file, and all with following format:

host_name:/vol/v1/v2

I only wanted to extract "host_name" and "v1" from every line in the file, then put these two words into a output file "out_file".

then create a for loop to process these two words in "out_file".

can anybody please help me out, it'd be appreciated very much

Thank you!

[Dennis Handly]

You could do:

IFS_save=$IFS # save
IFS=":/"
while read name dummy vol v1 v2; do
   echo $name $v1
done < in_file > out_file
IFS=$IFS_IFS   # restore

>then create a for loop to process these two words in "out_file".

You could process these in the above while.  Or you can create another while read with just the two fields.

No need to fiddle with IFS for this second while.

[Hanry Zhou]

Thank you all for the response. I have one follow-up.

while read filer path; do
/bin/scripx.sh -f $filer -v $path -a
done < ./output1

with this script, why in the end only one instance of /bin/scriptx.sh got excuted? I wanted to run same number of times as the number of lines in "output1" file.

Thank you again!

[Hanry Zhou]

First off, no, I did not put scriptx.sh under /bin, it is just an example. it is actually located somewhere else. the following script results in the same.

$ cat hv1.sh
#!/bin/sh

exec <./output1
while read filer path; do
./hv2.sh -f $filer -v $path -a
done

The cause of it is because that in-house script scriptx.sh (or ./hv2.sh in your case) takes quite some time to finish. I even put sleep 60 in the loop. It is still the same. I used exec < ./output1, it doesn't help.

output1 file is consists of multiple lines which contains filer-n and path-n, as you guessed.

[Hanry Zhou]

#cat run_find

exec <./output1
while read filer path; do
/example/find_details.pl -f $filer -v $path -a
sleep 60
done

#cat output1

filer1 path1

filer2 path2

filer3 path3

...

filer180 path180

sorry for the confusion. Let me try it again.

The script "find_details.pl" is the perl script and it is to grab some volume information from about 180 NAS storages. It takes about > 1 mins to complete.

The issue is the script "run_find" only execute the "find_details.pl" on the first line of "output1", then exit out.

Hope I made myself clear this time. thank you very much for your patience.

[Dennis Handly · Accepted solution]

.../scripx.sh -f $filer -v $path -a
>why in the end only one instance of .../scriptx.sh got executed?

If your evil scriptx is also reading stdin, you'll lose input.  You can fix this by:

.../scripx.sh -f $filer -v $path -a < /dev/null

Typically examples of this failing is use of telnet, ssh, remsh (without -n), etc.  Or scummy C shell scripts.

http://h30499.www3.hp.com/t5/Languages-and-Scripting/telnet-script-while-read/m-p/5824797

[Hanry Zhou]

This is the answer I am looking for. Thank you !