Structural Metadata is information about how the data should be structured in terms of its dimensionality, the concepts being measured and how they are classified. SDMX provides a range of 'structural metadata artefacts' including Concept (e.g.,frequency, country, gender), Codelist (classifications) and Data Structure Definition (describing the dimensionality of a particular dataset). Statisticians use structural metadata artefacts to define their statistical data model in Fusion Registry in the same way that a SQL database administrator will use tables, columns and keys to define the schema for a relational database.
While a Data Structure Definition (DSD) sets out the structure or dimensionality of a dataset, a Dataflow describes a specific instance. For instance, a DSD for National Accounts could be used for GDP, Central Government Debt and CPI Dataflows - different datasets that all share the same dimensionality. Data consumers retrieve the data they are interested in by querying the relevant Dataflow. In some ways, Dataflows are equivalent to Tables in the SQL relational model.
Structural metadata is also used to describe rules to be applied when collecting and processing data. In particular, Content Constraints set what data a provider is authorised to report.