Unified Medical Language System® (UMLS®)
UMLS Database Query Diagrams
The UMLS RRF (Rich Release Format) files are a collection of relational files containing the UMLS Metathesaurus data. The data in these files have been denormalized in a variety of ways to allow them to be more easily accessed and to retain the access paths similar to the Original Release Format (ORF). One consequence of this is an inability to render the data model as a strict set of Entity-Relationship (ER) diagrams. For example, the MRSAT.RRF file contains attributes on both atom objects (MRCONSO.RRF data) and relationship objects (MRREL.RRF data). Thus, the METAUI field of MRSAT.RRF links to both MRCONSO.RRF (AUI) and MRREL.RRF (RUI) files.
As an alternative, there are a set of ER-like diagrams called “query diagrams.” Each diagram considers a particular use case and shows how the files can be joined (or queried) to find all relevant data for that use case. The diagrams and accompanying queries assume a load of the RRF data files into an Oracle database using the load scripts created by the MetamorphoSys (MMSYS) installation tool. Another way to think about these diagrams: each one shows how to answer a particular kind of question involving UMLS Metathesaurus data. For example, the Source Concept (SCUI value) diagram shows how the tables can be queried and joined to look up all data associated with a SCUI. For each case, there are two parts:
- A diagram showing how the UMLS Metathesaurus tables can be joined together in the context of this use case. This is the “query diagram.”
- A set of queries (runnable against an Oracle load of RRF data) that enact the joins shown in the query diagram.
Note: the diagrams and queries are not indended to be complete or comprehensive. There are other aspects of Metathesaurus data that are not covered here. This is an attempt to provide information on the most common and most useful ways of interacting with UMLS data in the Rich Release Format (RRF).