Need help with data driven category

[ben]

Hi @all,

i am having trouble with a data driven category, when using the option "detect hierarchies".

Level 1:
   Tag: XMP::iptcCore\CountryCode\CountryCode\0
   Use other element = No
   "Rest" = default
   
Level 2:
   Tag: photools.com::IMatch\101400\file.foldername\0
   Use other element = No
   Detect hierarchie = Yes / No          -> see cases below
   Separators: "\"
   "Rest" = default   

Case 1 - OK:
   Detect hierarchie = No
   -> i get a tree structure, where all files are contained, that have the tag CountryCode not empty -> 10x
   
Case 2 - NOK?!
   Detect hierarchie = Yes
   -> i get a tree structure with ~13.000 elements   
   -> why are there images included that have no CountryCode set?
   
   
Thanks for any help,

Ben
   

[Mario]

I actually had to re-create this and to think about it for a while to figure out a) what you did and b) if IMatch is right. IMatch is right.

When you split folders into hierarchies in a data-driven category (which is unusual, all by itself) the top-node becomes the drive (e.g. c:). Below that, all folders from drive C: are added as child categories because of your split folder into hierarchy. And when IMatch determines if it must add "c:" below the country "USA" it checks if at least one file in "USA" is also contained on "C:". If this is the case, "C:" is added, and naturally all it's child categories, recursively.

I'm not sure if you can achieve what you probably are trying to do this way.

[ben]

I'm not sure if you can achieve what you probably are trying to do this way.

What i am trying to do is...

"At the end, i want a subset of my database to be synchronized to a linux device which can be accessed by FTP.
I want a very comfortable way for selecting this subset in iMatch."

So, i was thinking of creating a collection (data-driven based on a specific keyword is only one possibility)
and then use an iMatch script to sync it to the other device.

If i create this collection, using the same tree structure as the folder structure itself, then it is also
pretty easy to get a quick overview, which images are in this subset collection. That's why i tried this data-driven category
with splitted folders.

Mario, do you have some other suggestions?

Thanks for any help.
ben