Yep, we could talk about the details through another channel. However, I wanted to bring this little weird “factoid” to your reader’s attention, and maybe they could verify it themselves:

You define something such as /pageblabla.asp (Scenario step, URL Param page, etc.) and run it. If you get “0″ result, try again with checking the regex box (and escape \.asp; or you know what? you may not even need to specify the extension in the first place) and run the test again. Boom! You get results. It seems that WT “sees” the data only when regex is used in some circumstances.

OK, I’m pretty sure I haven’t dreamt this, but I will test it again, and let you know. Would be great if other readers tried it too.

The Test button is a little weird. It’s sensitive to capitalization although capitalization won’t affect matching in WebTrends itself. In a really complicated regex, I wonder if there are other quirks. I’ve never seen a regex engine that didn’t have something odd about it, sooner or later. In TextPad, for example, the backslash escape character combined with parentheses turns the parentheses into operators, when the “rules” say the \ should preserve the non-operator status instead.

This complicated regex talk is scaring me.

Can you send me the details of your latest finding?

However, something still puzzles me with the Reg Ex… checkbox ! I have noticed over the years that some types of reports just work better, I mean often work at all, if I choose the regex approach (Scenario Analysis steps, URL Parameter page, some others) !?

I got a case just today where defining a Content Group with regex showed no results (and yes, I tested the structure with the lovely and useful “Test” button, as well as tested the targeted URL), whereas not resorting to regex did… !

Yes, Regexes are lovely and useful; it’s just how WebTrends handles them in some circumstances that is not always clear.

But maybe it’s just me, as my wife would say…