The next question is how to make the data look relational when we actually want to view it as a set of related entities? Conventional relational database management systems (RDBMS) lack the means to efficiently access data inside the nested objects. Using the discoveryHub platform in conjunction with a modern RDBMS, we can construct an elegant solution to the problem.
The nested relational structure shown in diagram #2 creates a single relational table that stores a set of complex objects in a compressed format; for example, as a set of “large objects” in Oracle (BLOB/CLOG/LONG). To complete the query processing, we need to index the data and we also need the ability to ‘see inside’ the large objects. We need only expose the index-able fields (accession, sequence, etc) to the relational database management system: these become key fields and thus can be indexed. In this way, indexing is kept simple and efficient. The large object contains the complete structure in native form. To access the information within the large objects, discoveryHub is used to implement a large object parsing level that enables the query to see inside the compressed data structure. Because discoveryHub is designed to deal with nested structures, the parsing level can be implemented simply and efficiently, without degrading performance on large volume queries.
The resulting solution is powerful and scalable. discoveryHub provides an efficient platform for creating, updating and accessing a hybrid object/relational database. The platform provides the means to work with th