It sounds like you might be attempting to solve the wrong problem. Why are you interested in the differences between each pair of patents? If it's to find "similar" patents to a specified one, it might be faster to compute the results as needed. I can't really imagine that you would care about the difference between a patent on 'improved grooming proceedure for cocker spaniels' and 'increasing yield of aspartame using industrial waste'.
Once you identify why you need the results, you might find ways of getting what you need quickly without exhaustively building a database for every possible combination (6.4 x 10^13 entries ... sounds like a reasonably large database when you're done...). For example, if there's a threshold above which you don't care, you might be able to cluster your patents into groups such that membership in one group excludes it from all other groups, drastically reducing your search space. If you can do it well enough, you may be able to run queries against the data in real time.
...roboticusIn reply to Re: Huge data file and looping best practices
by roboticus
in thread Huge data file and looping best practices
by carillonator
For: | Use: | ||
& | & | ||
< | < | ||
> | > | ||
[ | [ | ||
] | ] |