[tasogare]

It depends on how the graph is stored in the database. In this project the nodes ids are TEXT so it will likely not scale very well. I know because I use a similar implementation with GUID as string in Sqlite in a project since a couple of years and while it works fine for the graph I have (<1 million nodes, few edges per nodes) it won’t perform too well past that.

[alexisread]

To some extent I think it depends on what data you're storing in the graph ie. If it's temporal data using a ulid instead of a guid speeds things up significantly (30x for large data) as your ids are not as fragmented.

https://github.com/schinckel/ulid-postgres/blob/master/ulid....

[mosselman]

Thanks for the info. Do you happen to have some stats to share?