Can textpipe do this IIS web server logfile splitting job
Posted: Mon Dec 08, 2003 9:27 am
I've got a slightly unusual web server logfile analysis job, I wonder if Textpipe can help.
I want to only keep the logfile lines that represent a 'first visit' from some host or IP address. AIUI this is slightly non-trivial (one needs to specify parameters like time lapse before a same IP address is considered a 'new visit' etc.)
As regards why I can't just use some standard web log analysis software...
I then have to run these logfile lines though Geolyzer from Geobytes which converts IP addresses to countries/locations (where possible - I know there are limitations to this mapping).
Finally I can import that lot into Excel and get some charts approximating the geographic distribution of visitors. (If I do this on the raw logfiles, I'm going to get useless results based on 'hits' - i.e. pages & images, rather than visitors.)
As a kludge, I could zap out request lines that aren't for .htm / .html / .asp - I'm sure Textpipe could do that, but doing so would skew the results if (say) Australian visitors looked at more pages than UK ones.
I want to only keep the logfile lines that represent a 'first visit' from some host or IP address. AIUI this is slightly non-trivial (one needs to specify parameters like time lapse before a same IP address is considered a 'new visit' etc.)
As regards why I can't just use some standard web log analysis software...
I then have to run these logfile lines though Geolyzer from Geobytes which converts IP addresses to countries/locations (where possible - I know there are limitations to this mapping).
Finally I can import that lot into Excel and get some charts approximating the geographic distribution of visitors. (If I do this on the raw logfiles, I'm going to get useless results based on 'hits' - i.e. pages & images, rather than visitors.)
As a kludge, I could zap out request lines that aren't for .htm / .html / .asp - I'm sure Textpipe could do that, but doing so would skew the results if (say) Australian visitors looked at more pages than UK ones.