Hi guys, hopefully this hasn't been asked before - couldn't see the question anywhere.

I have a large number of timestamps (hh-mm-ss-millisecond)
that I need to find the difference between e.g.:

14-11-07-513
14-11-07-644

Now the script that I have just knocked up is horrifically slow, the crux of which is converting the hh, mm, dd into milliseconds:

(( TOTALTIMESTART= (`echo $var | cut -d'-' -f4-8 | cut -d'<' -f1 | cut -c1,2` * 3600000) + \
(`echo $var | cut -d'-' -f4-8 | cut -d'<' -f1 | cut -c4,5` * 60000) + \
(`echo $var | cut -d'-' -f4-8 | cut -d'<' -f1 | cut -c7,8` * 1000) + \
(`echo $var | cut -d'-' -f4-8 | cut -d'<' -f1 | cut -c10,11,12`) ))

Once done on two values it subtracts them from each other and usese the difference.

There is obviously a better way of doing the part above, but I can't think for the life of me what it would be.

Any thoughts?

$ time ( echo "14-11-07-513" | awk -F- '{ print $1*3600000+$2*60000+$3*1000+$4 }' )
51067513

real    0m0.005s
user    0m0.004s
sys     0m0.000s

Wow, thanks for that. Exactly what I needed.

#!/usr/bin/perl
@arr=("14-11-07-513","14-11-07-644");
for($i=0;$i<=$#arr;$i++){
	my @tmp=split("-",$arr[$i]);
	$brr[$i]=$tmp[0]*3600000+$tmp[1]*3600+$tmp[2]*60+$tmp[3];
	print $brr[$i],"--->\n";
}
print $brr[1]-$brr[0];

Hi guys, well I was still having speed problem (albeit it was taking about 10 seconds to process the file instead of 50 seconds - it is a really big file) because both solutions required me to use a cut to fill the variables which obviously slowed it down a fair bit. I did a bit more digging around and thought I would share what I have come up with instead which takes less than 1 second to complete:

while read monkey
do
   set x $monkey
   TOTALTIMESTART=$(($((${2}*36000000))+$((${3}*60000))+$((${4}*1000))+$5))
   TOTALTIMEEND=$(($((${6}*36000000))+$((${7}*60000))+$((${8}*1000))+$9))
   (( TIMEVAR=($TOTALTIMEEND-$TOTALTIMESTART) ))
   do some stuff
done < $1