(8) Structural Metadata (V2)
- Charlotte Reynolds
What is “Structural Metadata”?
Structural metadata describes the variables within a dataset. This was previously referred to as Technical Metadata.
What do I need to include?
We recommend sites to provide a data dictionary of structural metadata that includes the following information:
Metadata Field | Completion Guidance | Constraints | Mandatory? |
|---|---|---|---|
Table Name * | Name of the table in the dataset. Use a fully qualified name if appropriate. | 255 characters max | Yes |
Table Description | Description of the table in the dataset. | 20,000 characters max | No |
Column Name * | Name of the column in the table dataset | 255 characters max | Yes |
Column Description | Description of the column in the table dataset. | 20,000 characters max | No |
Data Type Name * | Type of data contained in the column. | 255 characters max | Yes |
Sensitive * | Please indicate (True / False) whether the information must be treated as sensitive and may need additional constraints / removal / anonymisation / masking through the data access request process. Definition: An ODRL conformant policy expressing the rights associated with the resource. | True / False only. | Yes |
Where can I see some concrete examples of valid Structural Metadata?
Please see the following example for your reference. If you require further clarification, we also have several fully populated submissions from other sites that we would be happy to walk through with you. Please contact us via Service Desk to arrange a session.
Structural Metadata Example | |||||
|---|---|---|---|---|---|
Table Name | Table Description | Column Name | Column Description | Data Type | Sensitive |
Summary | The summary | Identifier | The identifier | String | TRUE |
Summary | The summary | Title | The title | String | FALSE |
Summary | The summary | Description | The description | String | TRUE |
Required | The required elements | Release Date | The release date | Date | FALSE |
Required | The required elements | Data Controller | The controller | String | FALSE |
Required | The required elements | Data Processor | The data processor | String | TRUE |
Required | The required elements | License | The license | String | FALSE |
Recommended | The recommended elements | Periodicity | The periodicity | Integer | TRUE |
Recommended | The recommended elements | Frequency | The frequency | String | FALSE |
What format should this be in?
The standard format required is an excel file which includes the Metadata Fields listed in the table above (Table Name, Table Description, Column Name, Column Description, Data Type, Sensitive).
Please see this template file:
Is an Excel spreadsheet the only option in which I can upload metadata?
Definitely not! The simple excel template is used as the low common denominator that makes our "lite" user interface as accessible as possible.
However, there are several different methods that supervisors can use to consume and update metadata. We've already used a few of these different formats, on request, with various other members:
REST API that accepts well-formed XML:
2. We can also accept reports from a profiling tool called white rabbit (https://www.ohdsi.org/analytic-tools/whiterabbit-for-etl-design/). White Rabbit is a .jar (with a UI if you need) that you can point at a DB and generate a profiling report from, which can then be used to create a detailed data dictionary.
If you would like to know more about them please raise a Service Desk ticket and we'd be happy to arrange a call to walk through some of them.