The Basics
The question is, why do you need a site access log? The answer is easy
- to see who visits your site. If you are a site owner and you have
a hosted web site, you probably have seen web stats of some kind compiled
by your hosting company. How do they do it? Often, they do it by analyzing
your site's access log.
A site access log is a record of all access to your site. Basically
it is a text file and each line is a record of a visitor's viewing of
a particular page. Typical site access logs include the name of the
page accessed, a user's IP address, the time of viewing, and some other
useful information.
By looking at a site access log, you can see how a visitor moves through
your site. You can see how many seconds it takes for a visitor to go
from one page to another. You can see which pages are navigated to and
which ones are avoided.
The Script
This handy site access log PHP code will create one text file for each
day of logging in a folder called logs. Note: you have to make a "logs"
folder in your base directory for this to work. The directory is not
created automatically. Place this code in any page you want to be recorded
in the log file.
Script
Notes
You can setup your $msg variable anyway you want to, but careful thought
should be done as to how you are going to parse it later. Inserting
a common delimiter, such as a symbol between elements of the log entry
can make it easier to parse later. For example, examine the following
$msg assignment:
If we had done this, we could use PHP's explode() or split() functions
and separate the elements easily, by splitting the line by the character
##. Why did I use two # and not one? Both the referrer and the query
variables have the possibility of having the # character in them because
this is used when defining an html anchor. Having two ## is still possible
though, however unlikely, so it would be a better idea to split them
with a random string that you'll probably never see naturally, such
as *%^%*.
You could also make your log files in the common logfile format, which
is supported by all major servers. A benefit of using this is that there
are downloadable scripts and programs that can analyze your logs for
you in you use this format. The definition of the format is on the W3C
site here.
I prefer to to write my own log analysis scripts however and prefer
to not use the common log format.
You'll also notice that we have put the @ symbol before the call to
the fwrite(), fopen(), and fclose() functions. This is the error suppressor
symbol, and it will prevent PHP from displaying any errors related to
these function calls. Because logging is likely secondary to the main
goal of our page, which is to provide the viewer with some sort of service
or information, showing them a failed error message is unnecessary and
unprofessional. Sure it works now, but if you delete the logs folder
at some point in the future, then you will see a nasty error displayed
on all pages containing this script. Instead of making all users see
this hideous error message, just check occasionally to see if any new
log files are being created. We also could have checked to see if the
folder exists before calling those functions, but it is difficult to
test every single possible point of error, and costly to do it every
time a page is requested.
When creating the log file, we use the format: date("Y_m_d")
. It is important to start your log file name with the year and not
the month or day. If we started the name with the month, a year from
now our logs would no longer be naturally sortable by name because the
file system will put all the December (12) entries together and all
the June (6) entries together. Instead, start the name of the file with
the year, followed by the month, and then the date, to allow for easy
natural date sorting by filename.
Now
What
Now that you have a text file full of page access records you can do
quite a few things. You can parse the info and gather the number of
unique IP addresses that visited your site. You can calculate the average
time per visitor, or number of page views per IP address. The details
are beyond the scope of this article, but very easily doable with a
little PHP tinkering.
Related Examples: Simple PHP Counter,
Advanced Object Oriented PHP Counter
|
