From a perfromence point of view is a fast full scan (ffs) of an index faster than a index range scan when they both read the index fully? I am trying to test this but i having diffiuclty to get my plan to use a ffs even when hinted. Context: My btree index has only 2 values N and P and i dont use a bitmap as i can't rebuild after every load as the table has 225M rows in it. the index only has 7M rows as the rest of the values are null and a particular query needs to access all 7m rows that the index refers to. So both the ffs and the range scan read the same data. My question is am I wasting my time trying to get the plan to do a ffs of the index? -- //www.freelists.org/webpage/oracle-l