On Tuesday, November 30, 2004 11:56 AM [GMT], Wallace Nicoll <[EMAIL PROTECTED]> wrote:
> Hi, > > I'm using Analog to analyse outgoing web requests through a proxy > server, and daily logs in native format are around 200Mb. I analog > them and create a daily cache file (about 60Mb). However, I'm trying > to analyse a month's worth of data and it takes about 10 hours to run > (from midnight), which is now impacting on normal daily activities on > the box. > > I have most of the LOWMEM options set to 3, apart from FILE and USER > which are set to 2. The USER report is fine, but my question relates > to the FILE section. I realise analog is intended for analysing > incoming requests, so this is a slightly unusual use. > > My required output is to report total traffic going to each remote > website, as in ... > > reqs: pages: bytes: directory > -------: ------: -----------: --------- > 7882103: 499445: 60151240840: http:// > 54609: 4170: 1315170232: http://news.bbc.co.uk/ > 150548: 6645: 688841636: http://www.bbc.co.uk/ > 113084: 26: 655929238: http://ad.uk.doubleclick.net/ > 24205: 692: 510716781: http://by101fd.bay101.hotmail.msn.com/ > > It looks as though the cache file for each day contains a line for for > every URL on each site, when what I want in the cache file is the > hostname section of the URL. Any idea how this can be achieved? > Perhaps some format of rewrite? I archive the native log files so I > don't need the cachefile for reference. Off the top of my head, I'd look at changing your LOGFORMAT command to interpret the URL field as a referrer field, and then use the Referring Site report (REFSITE ON) to generate the report you're looking for. Aengus +------------------------------------------------------------------------ | TO UNSUBSCRIBE from this list: | http://lists.meer.net/mailman/listinfo/analog-help | | Usenet version: news://news.gmane.org/gmane.comp.web.analog.general | List archives: http://www.analog.cx/docs/mailing.html#listarchives +------------------------------------------------------------------------
