Data Structure Definition (DSD) - SDMX Web Services

Dataset Definition

Dataset Structure Definition (DSD) describes how information in a specific dataset is structured. Knowledge of the structure is important, because it allows to later filter out desired information very precisely based on criteria to limit specific dimensions. The SOAP SDMX 2.1 query (using "GetDataStructure") presented below is used to inquiry the web-service to retrieve the structure of the dataset the user is looking for (nama_10_gdp).

The same information may also be obtained with a help of the REST protocol:

The result of the above query contains the full dataset structure definition. To learn more about dataset definitions, consult the SDMX: Tutorials.

Now let's have a look at the resulting XML. It contains the following significant elements:

Measure list - the list of observed and aggregated values. Note the 'OBS_VALUE' identifier which is later used to identify XML elements holding measures.

Dimension list - each measure (aggregated value) is described by certain properties like data unit, geographical location, aggregation granularity (frequency). They are referred to as dimensions and are representing data cube dimensions of the data warehouse where the information is stored. Dimensions will later be used to limit information obtained through the web service. Note the 'FREQ' dimension and its corresponding 'CL_FREQ' code list. It is also important to know that the 'position' attribute is used when creating REST queries.

Code lists - code lists enumerate values to describe a single dimension. For instance the 'CL_FREQ' list describes the following codes: D-Daily, W-Weekly, Q-Quarterly, A-Annual, M-Monthly, H-Semi-annual. Each of them can be used to select only data with required aggregation level (granulation). Below is an excerpt from the resulting XML. It shows how 'CL_FREQ' code list is represented in the SDMX dataset definition. Other code lists are represented in a similar way.

The table here below details the dimensions obtained for the concerned dataset "nama_10_gdp". Later, this information will be used for data retrieval and to restrict ranges of required statistical information.

Dimension Code list Description
FREQ CL_FREQ Granularity of aggregation level like for instance Annual, Monthly, etc.
GEO CL_GEO Geographical location or locations for which observed values are aggregated. This can be a state like Luxembourg, Belgium or a group like EU28 etc.
NA_ITEM CL_NA_ITEM Statistical indicator measured.
TIME_PERIOD -- Time period for aggregated value of measure
UNIT CL_UNIT Data unit used to present data.


Descriptor Concepts - descriptor concept is a more general property. SDMX defines three types of 'descriptor concepts' - dimensions, measures, attributes. In the XML there are additional elements to describe 'concepts' like multilingual textual description and concept identifiers.