Originally posted by chrisrlong
Now.... you add back in another 26,512 rows, thereby expecting a total of 53,024 leaf rows, *but* you already deleted a bunch, so what if the index re-uses those deleted entries?

Assume, for example, that out of 13,450 empty leaf rows sitting around, say, oh, 9732 of them got re-used.
This is the part I don't understand. Since my index entries are all unique, I would expect that the index to re-use an entry UNLESS it had the exact same key value. For example, if I inserted a row with an ID=1, and deleted it, I would not expect that index row to be used again if I inserted ID=2. It should only get re-used if I insert another row with ID=1. (According to "Oracle Performance Tuning Tips & Techniques, Richard Niemiec" at least.)

I buy your explanation, I just don't understand why the index rows would get re-used.