Here is my setup for what it's worth in deciding on your own.
I import the access.log file into a MySQL database every night and then use
Crystal Reports to report on web usage. I run Squid on a Linux box but MySQL on
Windows 2k Server so your setup may vary.
First I reformat the log file to my liking with a VB hack:
***snip***
myLine = myLine.Replace(",", ";") 'remove any (,) to avoid
problems later
myFields = myLine.Split(" ")
myDate = CDate("01-01-1970 00:00:00") 'Initialize date at
Unix 0 point
'First field should be a Unix date - turn it into a date by
adding it as seconds to 1970
intDate = CInt(myFields(0))
myDate = myDate.AddSeconds(intDate)
myDate = myDate.AddHours(-5) 'timezone
myFields(0) = myDate.ToString("u")
'Put the line back together
newLine = Join(myFields, ",")
'Remove duplicate (,)
Dim replaceLine = newLine.Replace(",,", ",")
While replaceLine <> newLine
newLine = replaceLine
replaceLine = newLine.Replace(",,", ",")
End While
'Split the domain and the page into separate fields
Dim lastComma As Integer = 1
Dim domainEnd As Integer
For i As Int16 = 1 To 6
lastComma = newLine.IndexOf(",", lastComma + 1)
Next
domainEnd = newLine.IndexOf("/", lastComma)
If Mid(newLine, domainEnd + 1, 1) = "/" Then
domainEnd = newLine.IndexOf("/", domainEnd + 2)
End If
If domainEnd = -1 Or newLine.IndexOf(",", lastComma + 1) <
domainEnd Then
domainEnd = newLine.IndexOf(",", lastComma + 1)
End If
newLine = Left(newLine, lastComma) + Mid(newLine, lastComma +
1, domainEnd - lastComma) + "," + Mid(newLine, domainEnd + 1)
'Use lastDateReached (stored in a separate file) to make sure
we haven't already imported these fields
If lastDateReached < myDate Then
'Write out the line we have edited
outFile.WriteLine(newLine)
lastDateReached = myDate
End If
***snip***
Then I import the new file into MySQL:
"C:\Program Files\MySQL\MySQL Server 4.1\bin\mysqlimport" --user=*****
--fields-optionally-enclosed-by=""" --fields-terminated-by=,
--lines-terminated-by="\r\n" ***database***
\\***server***\***share***\***squid***access.log
Matthew Williams
-----Original Message-----
From: Michael St. Laurent [mailto:[EMAIL PROTECTED]
Sent: Friday, February 16, 2007 6:52 PM
To: Squid Users
Subject: [squid-users] Suggestion for log file analysis?
I'm looking for suggestions about what to use for log file analysis.
We need something that can be used to drill down into the particular
pages a particular user was accessing during a given stretch of time. A
web interface would be nice but is not a requirement.
Squidalyser looks like it's close. I can't see how to specify a
date/time range for the query though.
--
Michael St. Laurent
Hartwell Corporation
