Folks, since I got no response to my initial post, I am re-phrasing the question.
Our current strategy to recover a table (when its been mangled by a DBA, batch or user), is to recover the RMAN backup pieces to an alternate machine (takes 4 hours for 160 G ).
Then we restore and recover the entire database to a point-in-time (takes another 2.5 hours).
So by the time we are done exporting the table in question, its almost 7 hours!.
In this context I experimented with TSPITR and am quite comfortable with all parts of its working except where it imports the tablespace into the Recovery instance.
We can't rely on Flashback table since DDL change invalidates Flashback recovery
Please let me know if there is an alternate method to address my issue.