http://qs321.pair.com?node_id=1000798


in reply to Processing ~1 Trillion records

Can you give some background information of the problem at hand?

What database do you use? What does the query look like?

Restructuring a massive array and writing CSV files sounds costly. What pattern? How many CSV files?

Profiling a (small) subset of the data might give some indication. Did you time how long the various steps take?

In short: could you show some code please?