I had a query from an attendee of my full-text indexing session at TechEd US. He asked how he can find out which documents (or how many) still need to be processed. I did a little investigation on this and here's my best guess:
<WARNING: Undocumented and potentially just a guess!>
1. Query for the objectid of your full-text index. You can do this by:
select * from sys.fulltext_indexes
2. Open an admin connection to your system ie: connect to admin:SERVER instead of SERVER.
3. Query as follows:
select * from sys.fulltext_index_docidstatus_2105058535
(the number on the end needs to be your full text index's object id not mine :-))
From what I can see, this table seems to hold details of documents not yet processed and it gets cleaned up as documents are processed. This is an internal table that you can see via:
select * from sys.internal_tables
</WARNING: Undocumented and potentially just a guess!>
Hope that helps someone.