how to write this script?

[bluelake]

I have a file which reads like:
PKGINST: LGTOclnt
VERSION: 6.1.1.Build.238
PKGINST: NSCPcom
VERSION: 20.4.70,REV=1999.10.13.17.54
PKGINST: SMCcvs
VERSION: 1.11.5
PKGINST: SMCossh
VERSION: 3.4p1
PKGINST: SMCossl
VERSION: 0.9.6g

how can I write a script to convert it like:

PKGINST: LGTOclnt VERSION: 6.1.1.Build.238
PKGINST: NSCPcom VERSION: 20.4.70,REV=1999.10.13.17.54
PKGINST: SMCcvs VERSION: 1.11.5
PKGINST: SMCossh VERSION: 3.4p1
PKGINST: SMCossl VERSION: 0.9.6g

That is to say, to merge every two line to one line, Thank you.

There are only 10 types of people in the world: Those who understand binary, and those who don't

 

[SBSaikia]

Simple way:

say your raw data file is: file1
issue the command:
grep PKGINST file1 > file2
grep VERSION file1 > file3
paste file2 file3 > file4

Now file4 contains ur required format.

-sbsaikia

 

[guggach]

assumed the file looks exactly like you show,
try the old fantastic sed.
put this in sed-cmd:
{
N
s/\n/ /
}

then run: sed -f sed-cmd filename
sure, sed can more.

guggach

 

[Ygor]

You can use paste, e.g...
[tt]
paste -s -d"\t\n" file1
[/tt]
...or...
[tt]
paste - - < file1
[/tt]

 

[nawlej]

In Perl:

#!/usr/local/bin/perl -w
use strict;
my $output;
open(INFILE, "< /file/to/read/from");
open(OUTFILE, "> /file/to/write/to");

while(<INFILE>){
 chomp;
 if ($_ =~ /PKGINST/){
    $output = $_;
 }elsif($_ =~ /VERSION/){
    $output .= $_
    print OUTFILE $output."\n";
 }
}

___________________________________
[morse]--... ...--[/morse], Eric.

 

[bluelake]

Thank you all, With all of your hints, I also got an awk way:

awk '/PKGINST/ {printf $0"\t"}; /VERSION/ {print $0}' f1

There are only 10 types of people in the world: Those who understand binary, and those who don't