> If we are to go for the pickle route, why not use shelve? I don't know what > size of these databases we are expecting, but with shelve it is not necessary > to read the entire file to e.g. get the data for a specific object if we use > their name as the top dictionary key.
As Ali said, it's small, but when the python stats finally get in the tree, shelve is a good option for storing stats outputs. Nate _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
