It's hard to find a solution unless looking at the source tables. Could you please post the source CSV tables which should be transformed into the Sheet1?
It doesn't seem to be possible to use an attribute value for the Start Row under Start Position. Otherwise you might have been able to use that after having used a Counter for the previous csv rows and added some empty rows and thus gotten the Start Row for the second set of csv rows.
A workaround would be to pre-process the data in these steps:
1. Copy the attribute names of the first csv files to a new "header feature" and run it through a Counter set to Global. (This might seem strange, to count just one feature, but it's global so it'll make sense.)
2. Rename the attributes of the first csv file to simple column names: A, B, C and so on.
3. Put a Counter (also set to Global) on the rest of the features of the first csv file.
4. Repeat the three steps for the next csv file.
5. Put a Sorter right before the FeatureWriter.
Something like that. An ugly workaround, and I'm sure someone will give you three lines of python code that will solve it for you perfectly instead. :-)
