trimming perl array elements

homey · 02-17-2008, 09:00 AM

Hi,
I am using a perl script to read the dansguardian access.log and output to an html file which I email to myself and other admins.
The site field can be humungus and I trim it down abit so one doesn't have to use the scroll bar to view the rest of the page.
For example:

Code:

http://toolbar.google.com/service/update?as=tbie&version=4.0.1020.6156&os=big&hl=en&tbbrand=SUNA&sd=com&osver=5.1&ossp=2.0&browser=6.0.2900.2180&rlz=1T4SUNA_en___US205&id=457FF06587E1BE25D4AC192011FA87B26DEACsQYIO&ds=1&lurslt=105,5,105,5,105  GET 146

I can get the basic information from that field and the next one using this perl code. This work fine, I just wondered if ther is a better, more correct way of doing this?

Code:

	$site = ($logfile_fields[4]); $site =~ s/(.{90}).*/$1/;
	$reason = ($logfile_fields[5]); $reason =~ s/(.*)\:.*/$1/;

otheus · 02-17-2008, 10:37 AM

Quote:

Originally Posted by homey

This work fine, I just wondered if ther is a better, more correct way of doing this?

Code:

	$site = ($logfile_fields[4]); $site =~ s/(.{90}).*/$1/;
	$reason = ($logfile_fields[5]); $reason =~ s/(.*)\:.*/$1/;

This is probably more efficient, for what it's worth:

Code:

        ($site) = ($logfile_fields[4] =~ /^.{90}/);
        ($reason) = ($logfile_fields[5] =~ /^(.*?):/);

The difference would probably only be noticed over very large files, like you say are being processed.

osor · 02-17-2008, 11:41 AM

And for the first one, you can eliminate RE scanning by using substr. So you end up with:

Code:

	$site = substr($logfile_fields[4], 0, 89);
	($reason) = ($logfile_fields[5] =~ /^(.*?):/);

homey · 02-17-2008, 12:08 PM

Thank you!

The first part works very nicely and looks much better than what I had.
However, the reason line pops an error when I use your code.

Code:

	($reason) = ($logfile_fields[5] =~ /^(.*?):/);

Use of uninitialized value in concatenation (.) or string at ./test.pl line 83, <LOG_FILE> line 10697.

osor · 02-17-2008, 12:35 PM

Quote:

Originally Posted by homey

However, the reason line pops an error when I use your code.

Code:

	($reason) = ($logfile_fields[5] =~ /^(.*?):/);

Use of uninitialized value in concatenation (.) or string at ./test.pl line 83, <LOG_FILE> line 10697.

What perl version are you using? You can always get rid of the ? (greediness modifier), it’s not strictly necessary if you have only one colon.

Code:

	($reason) = ($logfile_fields[5] =~ /^(.*):/);

homey · 02-17-2008, 01:02 PM

Thanks,
perl -v
This is perl, v5.8.8 built for i386-linux-thread-multi

that still pops the error...
Use of uninitialized value in concatenation (.) or string at ./test.pl line 83, <LOG_FILE> line 10697.

Just to clarify, I want to show everything up to the : in that field. If no : is there, just print what is there.
For example:

Code:

GET 7
GET 0
*DENIED* Banned site
Banned extension
*DENIED* Banned Phrase found

from these snips...

http://www.streamaudio.com/stations/player/pages/newplayer/nowplay/adtrack.asp?type=Replacement&adid=74158&station=WHKO_FM  GET 7

http://ads1.msn.com/library/dap.js *DENIED* Banned site: ads1.msn.com GET 0

http://dl.google.com/toolbar/T4/data/en/big/4.0.1601.4978-big/GoogleNav.cab *DENIED* Banned extension: .cab GET 0

http://www.wlwt.com/index.html *DENIED* Banned Phrase found:  gator  GET 89562

http://oascentral.clearchannel.com/RealMedia/ads/adstream_sx.ads/wlw-am/home/@Top,Middle1,Left1!Top?_RM_HTML_LANDINGSITE_=www.700wlw.com *DENIED* Banned site: clearchannel.com GET 0

http://oascentral.clearchannel.com/RealMedia/ads/adstream_sx.ads/wlw-am/newsondemand/@Top,Middle1,Left1!Top?_RM_HTML_LANDINGSITE_=www.700wlw.com *DENIED* Banned site: clearchannel.com GET 0

osor · 02-17-2008, 01:24 PM

I see. I think the initial assumption was that there is always a colon. Anyway, you can do something like this:

Code:

($reason) = ($logfile_fields[5] =~ /^(.*?)(:|$)/);

But at this point, I am not sure how much faster it is from:

Code:

($reason = $logfile_fields[5]) =~ s/:.*//;

homey · 02-17-2008, 03:48 PM

Thank you, wish I'd thought of such a nice solution!

Code:

	$site = substr($logfile_fields[4],0,87);
	($reason = $logfile_fields[5]) =~ s/:.*//;