Systems and processes may access data in various graph structures, often including data models that do not partially or fully utilize relations such as hierarchical graphs. For example, software for processing this non-relational data identifies a data model that includes a plurality of nodes independent of relations, each non-root node storing an identifier of the particular node's parent. The software then processes the data model using a generic graph service that utilizes nodes and relations. In some cases, such data models may include hierarchical graph structures and one or more of the nodes may be a recursive node.