Depends on your data and how it needs to be modeled and used, imo. If it’s sparse and interconnected then obviously SQL makes a lot of sense. But for data that

Depends on your data and how it needs to be modeled and used, imo. If it’s sparse and interconnected then obviously SQL makes a lot of sense. But for data that is very document-like e.g. 10Ks you’d get from SEC EDGAR, then document based is suitable. I’ve found even with relational databases, sometimes you need to denormalize the data a bit to achieve performance benchmarks for complex queries, and there’s a tipping point where you may want to sync your data into something document-based like Elasticsearch to get improved search capabilities, too.