Wouldnt it be easiers if you pulled all discovered addresses from an existing db
then query for there balances
and filter out the ones < x
Problem solved. ?
It's fairly easy to pull this info from a full node. The full node I run is in a VM with 4GB RAM, and the method I'm familiar with (
https://github.com/znort987/blockparser) requires a machine with closer to 128GB.