DBMS_SQL array processing - ORACLE 8.0.6

Question

It's been a while since I've used this.
DEFINE_ARRAY - the last parameter is "lower bound", which is the starting point. If it's set to 1, rows will be placed in the array starting with 1 and ending with the last row fetched. The entire result set from the EXECUTE appears to be loaded into memory using the array. Fetch 1 will retrieve rows 1 - 10, and put them in array loc. 1 - 10, next fetch will return 11 - 20 and put them in array loc 11-20.
I don't remember it being this way. Why can I not reset the array, load 10 at a time in array locations 1-10, and re-use 1-10 for each fetch? Will I not eventually blow something up if too much is loaded into the array?
Also, why would I want to set lower bound to anything other than 1? Why would I want to use -10 to 9, or 0 to 20, or other ranges? What purpose does this capability serve?
(Of course, back then I used host arrays in PRO*COBOL. It must have been different.)
Thanks.
KL

Answer 1

If you do not care about the rows fetched from a prior call to FETCH, you can do a .DELETE on the collection to remove any existing elements and thus freeing the memory.
The index positions will still start from the next available index position, so you should use .FIRST and .LAST on the collection.
If you have a set of queries that you run one after the other and these queries produce the end result in the same format as the previous, then you may want to keep the previous rows in the collection, and start from the next available index.