On Tue, Feb 7, 2023 at 3:47 PM Merlin Moncure <mmonc...@gmail.com> wrote:
> On Mon, Feb 6, 2023 at 1:22 PM Peter J. Holzer <hjp-pg...@hjp.at> wrote: > >> On 2023-02-06 20:04:39 +0100, Julian Backes wrote: >> But UUIDs are random and that plays havoc with locality. > > > This is really key. [...] the databases I've seen that are written with > the > UUID pattern appear to be written by developers oblivious to this fact. > Well, perhaps these developers are not dealing with temporally clustered data, like commerce related DB, and more scientific data? In any case, this developer will definitely investigate ULIDs, vs UUIDs, vs Ints (sequences), based on the info from this thread. I'm aware of fragmentation issues, and cache hit/miss issues etc, in general; but was probably not sufficiently making the mental connection with UUIDs and PostgreSQL. So thanks everyone.
