I keep pondering the performance chart, regarding uuid insert, shown towards the bottom of the article. I believe he was doing that test with PostgreSQL.
My understanding is that the performance is degrading because he has a btree primary key index. Is it possible to try a hash index or some other index type for a uuid primary key that would mitigate the performance issue he is recording?
After all, I can't think of any use case where I query for a "range" of uuid values. They are always exact matches. So a hash index would possibly be a really good fit.
I have many tables, several with more than 1 billion rows, that use uuid's as the primary key. Many of those uuid's are generated off system, so I can't play around with the uuid generation algorithm like he was doing.
I'm hoping to move to PG 10 any day now, and can migrate the data with updated index definitions if it will actually help performance in any way. (I'm always looking for ways to tweak the performance for the better any chance I get.)
Hash indexes unfortunately don't support UNIQUE indexes. At least not yet. So while you can use them for regular indexing, they cannot be used as a PRIMARY KEY.
