Recently I had a need to examine IIS log files - lots of them - to do some usage analysis. Analog has been my tool of choice, and although it takes a bit of learning, when coupled with the right config settings will display helpful tables, charts and graphs for summary-type information (click for a larger picture):
But when it came time to answer some more directed questions about web site usage, I thought the best way would be to import the IIS log files covering 2 years worth of data into SQL Server, and utilise my existing knowledge of T-SQL. There were some pointers towards using DTS, but it seemed to me that the functionality of hand-coding import routines, allowing for header records in the log files, and coding looping constructs was already taken care of in Microsoft's LogParser.
LogParser allowed me to import the IIS log files with no problems, it cost nothing, and with a bit of help on command-line parameters from Serge van den Oever's blog, I had the table created and the import running in no time (well, it took just over an hour to upload nearly 700 web log files totalling 920MB into my test server, so that's not really "no time"). There's also an unofficial support site at http://www.logparser.com/.
Once uploaded I used ordinary T-SQL to do my queries. Easy.
This may sound like a breathless schoolboy "you won't believe what I've just done" story. But LogParser really is brilliant :-)
Tags: sql server, logparser, IIS, web server