Unless you post a complete example of the line to be parsed,
we cannot help you. Since I am a nice guy, though, I will
explain why you need more than split for this problem.
The better solution would be to use the m// operator and
the grouping variables: $1, $2, etc. I'll explain by
parsing an entry from an Apache web server access log
$line = '127.0.0.1 - - [26/Mar/2001:16:01:07 -0500] "GET /stuff/ HTTP/
+1.0" 200 11874'
Each entry is seperated by dashes, brackets, or quotes -
but since we know the general layout, we can write a regualar
expression that is general enough to parse each line, but
specific enough to get the data we want - just the IP of
the referrer, the date stamp and the requested document (
along with request type)
use strict;
my ($ip,$date,$method,$file,$header,$status,$pid) = $line =~
/^([\d.]+) # $id = ip quad
\s*-\s*-\s* # skip over these
\[(.*?)\]\s" # $date = everything between the brackets
(\w+)\s* # $method = the method, usually GET or POST
([^\s]+)\s* # $file = everything UP TO the next white space
(.*)"\s* # $header = everything UP TO the next double quote
(\d+)\s* # $status = digits between spaces
(\d+)\s*$/x; # $pid = last set of digits
print "$ip\n$date\n$method\n$file\n$status\n$pid\n";
By no means I am an master of regular expressions, the ones
I chose just happen to work - there are better ways then using
.* - but a little badness won't kill ya' :)
Big Thanks to Albannach.
Jeff
R-R-R--R-R-R--R-R-R--R-R-R--R-R-R--
L-L--L-L--L-L--L-L--L-L--L-L--L-L--
| [reply] [d/l] [select] |
Enclose code in a <code> tag. | [reply] |