3 ms·
Every DB engine used at scale has a concept of snapshots and backups. This just looks like someone making a git-like porcelain for the same kind of DB manageme
by qbasic_forever 6y ago
Every DB engine used at scale has a concept of snapshots and backups. This just looks like someone making a git-like porcelain for the same kind of DB management constructs.
- zachmu 6y agoIt's not just snapshots though. Dolt actually stores the tables, rows, and commits in a Merkle DAG, like Git. So you get branch and merge. You can't do branch and merge with snapshots. (You also get the full git / github toolbox: push and pull, fork and clone, rebase, and most other git commands)
- qbasic_forever 6y agoYeah it's a neat idea but I struggle to think of good use-cases for merge, other than toy datasets. If I'm working on a service that's sharding millions of users across dozens of DB instances a merge is going to be incomprehensible to understand and reason about conflicts.
- MyNameIsFred 6y ago> Yeah it's a neat idea but I struggle to think of good use-cases [...] If I'm working on a service [...] I suspect that's simply not the use-case they're targeting. You're thinking of a database as simply the persistence component for your service, a means to an end. For you, the service/app/software is the thing you're trying to deliver. Where this looks useful is the cases where the data itself is the thing you're trying to deliver.