Rich Dempsey ([EMAIL PROTECTED]):
> Abstract: we want to remove a substring, starting with a semi-colon,
> and continuing either to the query string or to the end (whichever
> comes first), from the URL before Analog does anything with the
> request or the referrer. Can this be done in Analog, or is a helper
> required?
> We use ATG's product, Dynamo, for dynamic web content. To track the
> user's state, it attempts to set a sessionid cookie in the browser.
> For the very first page, before it knows if the browser sets
> cookies, and for subsequent pages if the browser doesn't set
> cookies, it will encode the sessionid cookie in all links on the
> page. The request as observed by the web server looks something
> like this:
> GET
>/US/en/corp/store/catalog/Product.jhtml;jsessionid=<cookie>?PRODID=6082&CATID=267
>HTTP/1.0
> where <cookie> is actually some undelimited string of alphanumeric
> characters. There are also pages without a query string that are
> cookied as well, for example:
> GET /;$sessionid$<cookie> HTTP/1.0
> Note for this different cookie, we don't have any equals sign, not
> that it matters much.
> These are rather useless for analysis, and explode the reports far
> larger than they need to be, not to mention than our beleagured
> disks can take. One might consider using an ALIAS construct to
> remove the offending substring, except that it seems to be applied
> _after_ Analog has already done considerable work on the log line.
FILEALIAS and REFALIAS occur on the data at the point it is read from
the logs. There are other ALIASes (REQALIAS and REFREPALIAS) that
process data at the reporting point. So using FILEALIAS should be
sufficiently fast.
> Further, an ALIAS to clobber the embedded cookie will preclude any
> subsequent ALIAS from being applied to the request/referrer.
This is an issue, bur you could use the report-level aliases to make
changes at the reporting stage. Or you could make more complicated
FILEALIAS and REFALIAS statements that do both the sessionid removal
and other aliasing in one pass (perhaps with REGEXP).
> We prefer to archive our log files unaltered, so if a preprocessor
> is required, how might we specify that?
You could pipe the logs through sed or perl to do that:
perl -pe 's/;[j\$]sessionid[=\$][^\?]+//' < logfile | analog ... -
--
Jeremy Wadsack
Wadsack-Allen Digital Group
+------------------------------------------------------------------------
| This is the analog-help mailing list. To unsubscribe from this
| mailing list, go to
| http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
| List archives are available at
| http://www.mail-archive.com/[email protected]/
| http://lists.isite.net/listgate/analog-help/archives/
| http://www.tallylist.com/archives/index.cfm/mlist.7
+------------------------------------------------------------------------
