Scalability of Indexes

Hi there,

Im exploring Graph databases for a SaaS based product. Sometimes in the product, we let the users create their own schema. In that case, do we have to run a job to create indexes? Or is there a way to query without indexing?


Unfortunately we don’t support querying without indexes as its not efficient. You can query with uid without indices but if you want to search on some other filed index is mandatory.

Thanks for the reply.

I think I must clarify here. The indices are only used to provide support for functions.

All other features, like traversal, joins, pagination, shortest path, etc. do not need any indexing. And when users would set their schemas, they would have to specify that they need indices explicitly for indexing to happen. Dgraph doesn’t rely upon indexing for operation.

@mrjn Does that mean searching on non-indexed predicates is possible, just inefficient?

This is pretty vital to our use-case, that predicates are searchable. Considering that we may have, potentially, millions of predicates – it seems searching non-indexed predicates is better than indexing all predicates.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.