3 ms·
I've still never been convinced by the use cases of graph databases. It just seems like a new coat of paint over NoSQL. I always hear vague claims of increased
by mjhay 2y ago
I've still never been convinced by the use cases of graph databases. It just seems like a new coat of paint over NoSQL. I always hear vague claims of increased performance (the same thing everyone said about NoSQL), without much backing it up. PG is just so performant and feature-rich these days, it's hard to beat.
- robertlagrant 2y ago"NoSQL" isn't a technology, though. It's an industry category. At that level of detail it'll all seem pretty confusing, I'm sure! Relational databases: decent at everything Document databases: insane at saving and retrieving trees of data; bad at other things Graph databases: insane at deep relational queries (e.g. give me all my friends of friends of friends); bad at other things Columnar databases: insane at columnar operations; bad at other things