I must have missed the original thread, but this one caught my eye. I have a script that parses firewall log files and counts the bytes that meet a certain criteria. Usually I do this in one month at a time and I occasionally get negative numbers as answers when I get up over a few billion bytes.
I started wondering if the way I was printing (I was using printf) had anything to do with this. I solved the problem by dividing all my lines by 1 million (I think I actually used 1024*1024), which makes the report more readable anyways. Instead of saying XXXX generated 3,453,452,235 bytes of traffic (which never worked anyways, what I did get was:) XXXX generated -2,245,321 bytes of traffic (which is obviously wrong). I get: XXXX generated 3,453 MB of traffic But now I'm worried about rounding errors...although preciseness isn't paramount, entries that generate much less than 1 MB now look goofy: XXXY generated 0.003 MB of traffic ...but that's a different problem. If anyone is really interested, I'll post the script...all 30 lines of it. -r -----Original Message----- From: Ben Siders [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 22, 2003 2:37 PM To: Christopher D. Lewis Cc: [EMAIL PROTECTED] Subject: Re: Size of number in scalar You'll be waiting a long time. Perl quickly moves into scientific notation and can handle arbitrarily large values. I wrote a similar program a while back and got bored with it when the count hit about 10^17. :) -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]