Dropping Large Indexes

Hi,

we all know to build large indexes using many fast methods like parallel,
nologging, etc.

how can i speed up dropping large local index partitions with sizes
exceeding 400GB? it takes a couple of hours on a 64-CPU IBM p595 box. that's
too much lost for me.

the mark index 'unsuable' trick doesn't help as my inserts/updates would
fail.

since it is not an index supporting a PK, i was thinking of dropping only
the index partition that i don't want (leaving the table partition intact).
i could later on build the index partition after the data load.

is that the only "fast" method?

thanks in advance,

regards
anand

---
All I need to make a comedy is a park, a policeman and a pretty girl -
Charlie Chaplin

Other related posts: