Gordon Hamm wrote:

> i actually tried to load the data into a memory table, but it was over 2 gigs..
> Maybe a simple array would be better..

My first reaction to this was: Why are you using a database for these
iterations? 8 million records of 10 fields (each taking 10 bytes) would
only constitute 800 MB of data. That fits into a large array quite
nicely these days. Therefore, I would read in the whole dataset, iterate
'till your hearts content, and then write out the whole shebang in the end.