Btw, if you really need high performance here use a key-value DB Or a real relational database configured with a "key" column and a "value" column and let the

Btw, if you really need high performance here use a key-value DB Or a real relational database configured with a "key" column and a "value" column and let the wonders of proper clustering and indexing provide you with the data in realtime no matter what load you put on it, because this dataset is shit tiny. No really. I'm dead serious. If you feel like optimizing this further, you can ofcourse also save memory by shifting countries and cities to separate tables, this reducing the size of the dataset noticeably, maybe by up to 40%. I'm just shaking my head at all the "database problems" MySQL users invent to excuse their choice of subpar DB.