Skip to Content
Last modified by Artur on 2025/09/10 11:19
From version 4.1
edited by Helena
on 2025/05/19 15:20
Change comment: There is no comment for this version
To version 3.7
edited by Helena
on 2025/05/19 15:18
Change comment: There is no comment for this version

Summary

Details

Page properties
Content
... ... @@ -42,7 +42,7 @@
42 42  
43 43  * **Reference Metadata**: In addition to describing and specifying data structures and formats (along with related structural metadata), the version 2.0 specification also provides for the exchange of metadata which is distinct from the structural metadata in the 1.0 version. This category includes “reference” metadata (regarding data quality, methodology, and similar types – it can be configured by the user to include whatever concepts require reporting); metadata related to data provisioning (release calendar information, description of the data and metadata provided, etc.); and metadata relevant to the exchange of categorization schemes.
44 44  * **SDMX Registry**: Provision is made in the 2.0 standard for standard communication with registry services, to support a data-sharing model of statistical exchange. These services include registration of data and metadata, querying of registered data and metadata, and subscription/notification.
45 -* **Structural Metadata**: The support for exchange of statistical data and related structural metadata has been expanded. Some support is provided for qualitative data; data cube structures are described; hierarchical code lists are supported; relationships between data structures can be expressed, providing support for extensibility of data structures; and the description of functional dependencies within cubes are supported.
45 +* **Structural Metadata**: The support for exchange of statistical data and related structural metadata has been expanded. Some support is provided for qualitative data; data cube structures are described; hierarchical code 98 lists are supported; relationships between data structures can be expressed, 99 providing support for extensibility of data structures; and the description of 100 functional dependencies within cubes are supported.
46 46  
47 47  The major changes from 2.0 to 2.1 can be briefly summarized:
48 48  
... ... @@ -113,7 +113,7 @@
113 113  
114 114  [[image:SDMX_2-1_SECTION_1_Framework_2020-07_b3f80ada.jpg||height="417" width="564"]]
115 115  
116 -**Figure 1: High Level Schematic of Major Artefacts in the SDMX Information Model**
116 +**Figure 1: High Level Schematic of Major Artefacts in the SDMX Information Model**
117 117  
118 118  == {{id name="_Toc38914"/}}3.4 The SDMX View of Statistical Exchange ==
119 119  
... ... @@ -232,7 +232,10 @@
232 232  While the SDMX-EDI format is primarily designed to support batch exchange, SDMXML supports a wider range of requirements. XML formats are used for many different types of automated processing, and thus must support more varied processing scenarios. That is why there are several types of messages available as SDMX-ML formats. Each is suited to support a specific set of processing requirements.
233 233  
234 234  1. //Structure Definition~:// All SDMX-ML message types share a common XML expression of the metadata needed to understand and process a data set or metadata set, and additional metadata about category schemes and organisations is included. Also, the structural aspects of data and metadata provision – dataflows and metadataflows – are described using this format.
235 -1. //Generic Data~:// All statistical data expressible in SDMX-ML can be marked up according to this data format, in agreement with the contents of a Structure Definition message. It is designed for any scenario where applications receiving the data need to process it according to a single format. Such applications may need independent access to the data set's structure before they process it. Data marked up in this format are not particularly compact, but they make easily available all aspects of the data set. This format does not provide strict validation between the data set and its structural definition using a generic XML parser. It supports the transmission of partial data sets (incremental updates) as well as whole data sets. It supports both the timeseries and the cross-sectional use cases.
235 +1. //Generic Data~:// All statistical data expressible in SDMX-ML can be marked up according to this data format, in agreement with the contents of a Structure Definition message. It is designed for any scenario where applications receiving the data need to process it according to a single format. Such applications may need independent access to the data set's structure before they process it. Data marked up in this format are not particularly compact, but they make easily available all aspects of the data set. This format does
236 +
237 +not provide strict validation between the data set and its structural definition using a generic XML parser. It supports the transmission of partial data sets (incremental updates) as well as whole data sets. It supports both the timeseries and the cross-sectional use cases.
238 +
236 236  1. //Structure-specific Data~:// This format is specific to the Data Structure Definition of the data set (in other terms, it is DSD-specific) and is created by following mappings between the metadata constructs defined in the Structure Definition message and the technical specification of the format. It supports the exchange of large data sets in XML format (typically the size of the data set is 50% of the same data expressed as Generic Data), provides strict validation of conformance with the DSD using a generic XML parser, and supports the transmission of partial data sets (incremental updates) as well as whole data sets. The Structure-specific Data format specified in SDMX 2.1 supports both the time-series and the cross-sectional use cases which were covered by two distinct formats in SDMX 2.0. Many XML tools and technologies have expectations about the functions performed by an XML schema, one of which is a very direct relationship between the XML constructs described in the XML schema and the tagged data in the XML instance. Strong data typing is also considered normal, supporting full validation of the tagged data. These message types are designed to support validation and other expected XML schema functions.
237 237  1. //Generic Metadata~:// All reference metadata expressible in SDMX-ML format can be marked up according to this schema. It performs only a minimum of validation, and is somewhat verbose, but it does support the creation of generic software tools and services for processing reference metadata.
238 238  1. //Structure-specific Metadata~:// For each metadata structure definition, an XML schema specific to that structure can be created, to perform validation on sets of reported metadata. This structure is less verbose than the Generic Metadata format, and, because the XML mark-up relates directly to the reported concepts, it is appropriate for applications that are designed to process a specific type of metadata report. It is analogous to the Structurespecific Data format for data in its approach to the use of XML.
... ... @@ -275,8 +275,10 @@
275 275  
276 276  = {{id name="_Toc38926"/}}10 Validation and Transformation Language =
277 277  
278 -For many years the SDMX initiative has been fostering and supporting the development of a standard calculation language, called Validation and Transformation Language (VTL). A blueprint for defining calculations was already described in the original SDMX 2.1 specifications (package 13 of the Information Model - “Transformations and Expressions”). It was just a basic framework that required further developments to became operational in order to achieve a calculation language able to manipulate SDMX artefacts.
281 +For many years the SDMX initiative has been fostering and supporting the development of a standard calculation language, called Validation and Transformation Language (VTL). A blueprint for defining calculations was already described in the original SDMX 2.1 specifications (package 13 of the Information
279 279  
283 +Model - “Transformations and Expressions”). It was just a basic framework that required further developments to became operational in order to achieve a calculation language able to manipulate SDMX artefacts.
284 +
280 280  These developments started in late 2012 and were put in charge of the Validation and Transformation Language Task Force (VTL TF), which included members of the SDMX Technical Working Group (TWG) and Statistical Working Group (SWG), besides experts coming from the DDI and GSIM communities. The intent was to define a stardard language to be implemented in SDMX and applicable also to GSIM and DDI. This brought to the publication of the VTL 1.0 in 2015. Then new requirements came from a number of proofs of concepts and tests of VTL 1.0 made by several organisations and triggered a large improvement of the language. A new provisional version, the VTL 1.1, was released in public consultation in 2017. The high number of comments received triggered another phase of intensive work, with the main goal of achieving a more robust and forward compatible version. Finally, the VTL 2.0 was published between April and July 2018 (see the SDMX website).
281 281  
282 282  The implementation in the SDMX standards of the VTL 2.0 started in late 2018. It allows to write VTL 2.0 programs for validating and transforming SDMX data, to store these programs in a SDMX metadata registry and to exchange them through SDMX messages, also together the definition of the data structures of the involved data.
... ... @@ -285,8 +285,10 @@
285 285  
286 286  The new SDMX 2.1 parts relevant to VTL are in Section 2 (Information Model), Section 3A (SDMX-ML: XML formats), Section 3B (SDMX-ML: XML Schema, Samples, WADL and WSDL), Section 5 (Registry Specifications), Section 6 (Technical Notes) and in the REST specifications of the SDMX web services. On the contrary, the VTL is not supported in the UN/EDIFACT SDMX formats (Section 4) and in the SDMX SOAP web service interface.
287 287  
288 -As for the Section 2 (Information Model), the change is relevant to the package 13 “Transformations and Expressions”, which previously contained only a blueprint for future elaboration and was not operational. This package is renamed to “Validation and Transformation Language” and is completely reformulated by introducing the actual object classes needed to define and manage VTL 2.0 programs. The other parts of the IM are not impacted.
293 +As for the Section 2 (Information Model), the change is relevant to the package 13
289 289  
295 +“Transformations and Expressions”, which previously contained only a blueprint for future elaboration and was not operational. This package is renamed to “Validation and Transformation Language” and is completely reformulated by introducing the actual object classes needed to define and manage VTL 2.0 programs. The other parts of the IM are not impacted.
296 +
290 290  In the Section 3A and 3B, new schemas, samples and documentation aimed at exchanging VTL programs through SDMX structure messages have been introduced. Other parts are not impacted.
291 291  
292 292  In the section 5 (Registry Specifications), the VTL artefacts have been added. The other artefacts are not impacted.