What is “Structural Metadata”?
Simply put, structural metadata describes the variables within a dataset and how they are organised.
In DSRC Data Standards this takes the form of a Data Dictionary.
The Data Dictionary is organised into Data Classes.
A Data Class, or a table, is a collection of one or more Data Elements and/or child Data Classes.
The definitions, including rules and constraints for valid values for all Data Elements are defined using the following metamodel:
Name: the name of the data class
Description: a description of the data class
Data Elements: data elements contained with the class
Child Classes: data classes contained within the class
Cardinality: the number of times a data class can occur within a group
Level: Could, Should, Must
Metadata: any additional metadata associated with the data class
Relationships: related data classes or elements (which may reside in another standard)
A Data Element is the logical definition of a unit of data such as:
a column in a database table
a field in a form
an element in an XML schema
The definitions, including rules and constraints for valid values for all Data Elements are defined using the following metamodel:
Name: the name of the data element
Description: a description of the data class
Data type: the constraint on the data element e.g. integer, string
Cardinality: the number of times an lement can occur within a group
Level: Could, Should, Must
Metadata: any additional metadata associated with the data element
Relationships: related data classes or elements (which may reside in another standard), including child/parent relationships
Data Elements contain a Data Type which specifies the input value of the element e.g. varchar or datetime.
Name: the name of the data type
Description: a description of the data type, if applicable
Type
Simple - e.g. varchar, string, integer
Enumerated - a list of possible values
Reference
Relationships: to other data types or data element
