Nothing to it really...

I build a token that looks at several possible matches and that token equals T if any one of the possibilities is correct. 

But it's running all of the things inside the conditional whether or not the conditional is satisfied. 

The conditional section shows everything set right and nested. (I created a new conditional and copied everything under this one, then disabled the original but this new one still runs everything.)

I have multiple other identifications set up identical to this. I like to setup a total token that uses a token input of: T(%token1)%(Token2)(%Token3)  where I return a pattern match of "(T).." if I want two of the tokens to match. That way I can add (%Token4) and (%Token5) later of I decide I need to match 2 out of 5 without having to rewrite all of the conditionals. This also makes it easy to do a "match at least X out of X!"  I just make sure each identifying pattern match token only returns one character. 

I did notice one thing: my rough total token comes back green when viewed through Output. All other lines are black. There's no errors or anything under Processing Information.

Do you have anything in the Pre-processing section?  Which token is being used for the "medline rough total" pattern matching?

The screenshots are sort of inconclusive. It's hard to tell where the tokens come from based on their names only. I'm not seeing an indication that Zone OCR ran in your screenshot. Could you post a screenshot of the Processing Info pane with the process names showing?

Miruna,

Here's one that should be there and works:

Here's one on the next page that should not work:

and you are right, I don't see the OCR zone in the processing pane. However it's still trying to run all those pattern matches and not skipping them like the other documents.

I figured out what causes it. I recreated the entire document class from scratch and got the same issue. I then created a test class that has a pattern match that will always fail and a corresponding invalid condition always that should never run. 

Inside the condition is a zone OCR and a pattern match with a token named "this should not run".

Testing this worked, the condition failed and the items inside the condition never ran. However when I added Line Removal to the Zone OCR it looks like it's executing a line removal on every single page and every single page now has the "This should not run" token showing up on it. 

Here's the settings:

(all default except vertical)

Pretty basic "identification" that should fail everytime and not run what's in the conditional. However as soon as I add the line removal it starts trying to do that and trying to do the pattern matching 2 process inside the conditional. Removing that off my problematic process "fixes" it there but the results are less than I'd like since I really do want to remove lines. For identification purposes it works without that sub process, I was just trying to save cpu processes in my page processing section by not having to zone OCR that area twice. 

I did update to the latest hotfix for QF as well.