DQL not a serious api. Lacking several key features.
Lots of other issues…
So what to do? Wouldn’t this be a good time to pivot?
Let’s be a bit realistic. The critical team member who built Badger has left. There’s no domain knowledge at Dgraph. What to do?
Pivot.
Ditch Badger and fork Cassandra like ScyllaDB has done. Instantly features like scalability both locally and geo are realised. Not only that, but ingest speeds are also increased.
Discord (before they moved to ScyllaDB), was using Cassandra. So it’s proven to scale. More so than Dgraph I would wager.
With CQL, there is already a base to draw from. Much like EdgeDB, you could either make a new api that uses CQL. Or perhaps draw from Gremlin?
Bottom line is. Dgraph has been failing on core issues. There’s an opportunity to be a commercial instance of JanusGraph. JanusGraph unfortunately is written in Java and is very difficult to setup. Wheres a newer Dgraph could be all written in Golang and easily setup and running in the enterprise.
Hey I get you point but I need a DB which handles relationships very well. The only viable option I found was DGraph. In case of relations I’ve got to make multiple queries to get what I want with other DB’s. There’s no other viable Graph DB alternative for me
I’ve heard of ScyllaDB but I’m not sure of moving to that from Dgraph as most of my queries will be based on relationships.
I’ve just tried to try Dgraph opensource and I’ve got to say I’m stuck.
I’m talking about Dgraph only. There’s no easier alternative for Dgraph for dealing with relationships in data(a graph db essentially)
Isn’t scylla a NoSql db?
