read the input file and write in the given format

suchi_s · 12-13-2004, 04:25 AM

i have input data in a file innthe format
1874500001 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N
/CLIP A Y N
1874500002 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N
/CLIP A Y N
1874500003 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N
1874500004 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N
1874500008 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N
1874500009 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N
1874500010 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N

i want the data in the format
1874500001 /CWHFE A N N /KWSUBT A Y N /OCBSUB A N N /CLIP
every line starting with ten digit number..
as in second line the data should be pribnted in one complete line..
not able to do..
some lines are in one comltete line... some are not.. how to deal

as i m doing

while read no1 no2
do

if [ ${#no1} -eq 10 ]
then
#echo ${#no1}
echo -n $no1 $no2
else

echo $no1 $no2
fi
done < outty

but those which are already in one complete line.. they start appending

Cedrik · 12-13-2004, 07:50 AM

Do you want just append "/CWHFE A N N /KWSUBT A Y N /OCBSUB A N N /CLIP" to the number
found in the beginning of each line ?

Or do you want concatenate /CLIP A Y N to the previous line to make it one line data ?

suchi_s · 12-13-2004, 10:12 PM

concatenate /CLIP A Y N to the previous line to make it one line data

suchi_s · 12-14-2004, 03:03 AM

concatenate /CLIP A Y N to the previous line to make it one line data

blackzone · 12-14-2004, 04:33 AM

if it's not a big file. and you don't mind doing it using a stupid way.

just write a c function which read one line at a time and store it in buffer

if it don't start with /CLIP, strcat "\n" and the line

else strcat " " and the line

after you finish save the buffer to file.

suchi_s · 12-14-2004, 04:36 AM

it is a very big file..
this wont be efficient..
anything in shell scripting

bigearsbilly · 12-14-2004, 06:12 AM

this mess seems to work.
Give it a good test though, especially first and last lines.

Code:

sed -n '

$p
/^[0-9]/ {
            H
         }

/^\/CLIP/ {
           H
           g
           s/^\n//
           s/\n\/CLIP/ \/CLIP/g
           p
           s/.*//
           x
           d
            }

it's in three parts.

$p means print the last line
the H one means if line starts with a digit append it to the 'hold' space (newlines are included)
last long messy bit says,
remove first newline,
remove newlines before '/CLIP'
print it out
delete the line excchange with hold space, delete line (long way of clearing the memoryfor next lines)

regards billy

bigearsbilly · 12-14-2004, 09:02 AM

Sorry, the last post was wrong!

Not sure how well it will cope with a HUGE file!

(I am at work so can't test fully!!!!)

Code:

cat YOUR_DATA_FILE |
sed -n '

/./ H

/^\/CLIP/ {
           s/.*//
           x
           s/^\n//
           s/\n\/CLIP/ \/CLIP/g
           p
           d
        }

$ {
    x
    s/^\n//
    p
}

for each line:
paste it to the hold memory ##### (the hold memory remembers the lines including the newlines )

Line starts with a '/CLIP' ?
paste it on memory, recall all memory, remove initial '\n' in hold (it adds one before pasting) remove \n before '/CLIP' in all cases
print it.
delete the line (forces read of next line)

last line, if anything in memory
remove the initail '\n' print

else do nothing

billy

chrism01 · 12-17-2004, 01:12 AM

Here's the Perl soln:

Code:

#!/usr/bin/perl -w
use strict;

my (
    $data_file,
    $data_rec
    );

$data_file = 'test.dat';
open( DATA_FILE, "<$data_file" ) or
                die "Can't open text file: $data_file: $!\n";

while ( defined ( $data_rec = <DATA_FILE> ) )
{
    chomp($data_rec);
    if( $data_rec !~ /^[0-9]/ )
    {
        print "$data_rec";
    }
    else
    {
        print "\n$data_rec";
    }
print "\n";
close (DATA_FILE) or
            die "Can't close input data file: $!\n";