The &store_line Subroutine - Page 9

We previously saw how the &store_line subroutine was being invoked to process each line from the log file that didn't represent an image request. Now let's skip down to the bottom of the script and see what that &store_line subroutine actually does:

# script proper ends. subroutines follow.

sub store_line {

  # store one line's worth of visit data

  my($host, $date, $time, $url, $referer, $agent) = @_;
  my $seconds = &get_seconds($date, $time);

  if ($visit_num{$host}) {
    # there is a visit currently "working" for this host
    my $visit_num = $visit_num{$host};
    my $elapsed = $seconds - $last_seconds{$visit_num};
    if (($expire_time) and ($elapsed > $expire_time)) {
      # this visit has expired, so start a new one
      &new_visit($host, $date, $time, $url, $seconds,
        $referer, $agent);
    } else {
      # this visit has not expired, so add to existing record
        &add_to_visit($host, $date, $time,
          $url, $seconds, $elapsed);
    }
  } else {
    # there is no visit currently "working" for this host
    &new_visit($host, $date, $time, $url, $seconds,
      $referer, $agent);
  }
}

Most of the Perl in this subroutine should look pretty familiar by now. In essence, this routine is functioning as a traffic cop, using the host and the time of this access to figure out if this request represents a new visit, or another request in a previously started visit. First, it checks to see if a key exists for the current line's $host in the %visit_num hash. If there is, it means we've previously processed a request from this host, so the script checks to see if the currently working visit for this host has "expired." That is, it looks to see if the time difference between this host's last access and the current access is greater than the value stored in the $expire_time configuration variable. If it is, it means enough time has gone by that this access needs to be considered the beginning of a new visit, and the script invokes the &new_visit subroutine. If it isn't, the script invokes the &add_to_visit subroutine instead. Finally, if there wasn't any key for the current $host in the %visit_num hash, if means this host hasn't been seen before at all. Accordingly, the &new_visit subroutine is invoked to create a new entry for it. Here's where we've implemented the feature of turning off visit expiration for cases where the $expire_time configuration variable has been set to 0. We've done that by making the logical test that determines whether a visit has ended actually contain two logical tests, both of which must be true for the "true" branch to be invoked:

if (($expire_time) and ($elapsed > $expire_time)) {

This works because joining two logical tests with and requires both of them to be true for the expression as a whole to be true. If $expire_time is set to 0, which is a false value, the test can never return true. So, again, three subroutines are invoked from within this &store_line subroutine: the &get_seconds subroutine, which accepts as arguments the date and time strings from the current log file line and converts them to something called Unix seconds. That routine, and the date arithmetic it performs, is the subject of the next chapter. The &new_visit and &add_to_visit routines, which handle the updating of the script's data structure, will be covered in Chapter 10.

Storing the Data - Page 7
Perl for Web Site Management