Date time problem

Hi Guys,
I have a file a.txt
Start Date/Time End Date/Time from Prob_Dura.
----------------- ----------------- ----- ------
20090525 23:58:59 20090526 00:00:00 machine1 000051
20090525 23:58:09 20090526 00:00:11 machine2 000150

The perl or shell script can:

1. remove first two lins.
2. remove column "from".
3. calculate how many seconds between End and Start Date/Time.
4. calculate how many seconds for the Prob_Dura, now the format is "hh/mm/ss".

The output should be:
61 51
122 110

I read the thread here, and know awk, sed can do, but when trying on my file, then still cannot figure out... ---

Thanks

put here your awk, even does not work. Let see it together.

Start Date/Time End Date/Time from Prob_Dura
----------------- ----------------- ----- ------
20090525 23:58:59 20090526 00:00:00 machine1 000051
20090525 23:58:09 20090526 00:00:11 machine22 000150

1. remove the first two lines.
   awk 'FNR>2' a.txt

2. remove the columun for from
   cut -f1,2,3,4,6 -d' ' a.txt
   but i fail to do that, becuase you can see the line no.2 machines22 + one space + 000150; but line no.1 is two spaces between the "from" and "Prob_Dura". So the -f6 is the empty character.

i stuck here.

Please advice, any other good way can select what i want, or solve the more than one space issue.

Thanks

Hi, if your day range is less than one month, i think below perl script is ok.

But if you day range may be more than one month, need to take diversity days-per-month into account and modify below code accordingly, which is a little bit complex:)

sub toSecond{
	my ($time)=@_;
	my @tmp=split(":",$time);
	return $tmp[0]*3600+$tmp[1]*60+$tmp[2];
}
while(<DATA>){
	next if $. < 3;
	chomp;
	my @tmp=split;
	$tmp[5]=~s/([0-9]{2})([0-9]{2})([0-9]{2})/$1*3600+$2*60+$3/e;
	print toSecond($tmp[3])-toSecond($tmp[1])+ ($tmp[2]-$tmp[0])*24*60*60," ",$tmp[5],"\n";
}
__DATA__
Start Date/Time End Date/Time from Prob_Dura.
----------------- ----------------- ----- ------
20090525 23:58:59 20090526 00:00:00 machine1 000051
20090525 23:58:09 20090526 00:00:11 machine2 000150

Hi cherry,

It works perfectly. I always thought awk, sed, those commands, but yours coding is so simple.

instead of
__DATA__
Start Date/Time End Date/Time from Prob_Dura.
----------------- ----------------- ----- ------
20090525 23:58:59 20090526 00:00:00 machine1 000051
20090525 23:58:09 20090526 00:00:11 machine2 000150

how i can do with a file, i mean the start date... 2009... 000051, those data in a "a.txt". how the perl script can read from a file, stead of __DATA__?

Thanks

-----Post Update-----

$data_file="a.txt";
open(DAT, $data_file) || die("Could not open file!");

while(<DAT>){