Interestingly, my previous iteration of the program did something very similar to File::Cache. I created a multi-level directory decimal tree indexed by the most-significant digits of the index. I then stored the values in the leafs of the lowest branches. The value could then be looked up by reading the corresponding file.
in reply to Re: Store large hashes more efficiently
in thread Store large hashes more efficiently
I also implemented some simple caching where a hash stored the most recently referenced values to avoid file accesses where possible.
However, I found the file accesses to be exceedingly slow (10x penalty) and the cache didn't help me much since there was little correlation between neighboring accesses. In fact, this is what led me to just store everything in a hash since the caching-via-filesystem-tree was too slow. And this is what has led me to try to optimize the storage size of the hash...
Still fascinating that I did essentially create my own File::Cache like approach before discarding it...