Hi,

 

 

Using the Dynamic Schema mechanism and the Fanout Dataset option would be a solution.   1) When adding MITAB reader, use wildcard in the Dataset to read all *.tab files under the root directory and its subdirectories. And also select 'Single Merged Feature Type'.

 

* Replace ? mark in the images with \\ (back slash). In Japanese language environment, a back slash is displayed as a yen mark on the screen.   2)  Expose 'fme_basename' attribute (i.e. file name) of the souce feature.

 

3) Add a DBF writer and set 'Dynamic Properties'.

 

  4) Set fanout dataset parameters of the DBF writer: i) Insert an AttributeSplitter transformer between the MITAB reader and the DBF writer to split fme_basename into a state part (e.g. 'CA') and a category part (e.g. 'ACC'). And expose the first element (i.e. the state part = _list{0}) of the list of split parts. ii) Double click or right click "<Output Root Folder> lDBF] > Parameters > Advanced > Fanout Dataset" node on the Navigator window to open 'Edit Fanout Dataset Parameters', and set parameters like this:

 

 

Hope this helps.

 

 

Takashi  

---

I'm not quite understanding step 4i. In the AttributeSplitter parameters, what do I enter in the 'Delimiter or Format String' field? 

---

Nevermind, I understand that field now.

---