4 ms·
Point is, ideally you want both. Especially master data is closely entangled with data which you have in tables (and rightly so). In the example above consider
by chopin 10y ago
Point is, ideally you want both. Especially master data is closely entangled with data which you have in tables (and rightly so). In the example above consider an order header table which will have a reference to an employee (the one working on the order or the one who created the order). The schema for orders is rightly relational but I'd need a join to the employee which therefore is stored in a relational table as well.
- hvidgaard 10y agoI don't see how that scenario is anything but relational, even for queries, or are you talking in extension of the previous post about the managerial line? If so I agree, but instead of "poluting" the RDBMS with graph related constructs, I think the option here would be save the data in the RDBMS, and either build a graph on demand, or maintain an in memory graph, for queries. "Everything but the kitchen sink" are rarely a good idea for tools, it has a tendency to make the cost of switching some module unacceptable.