ISSUE-37 (SDMX-2.1): Align Data Cube with SDMX 2.1 [Data Cube Vocabulary]

ISSUE-37 (SDMX-2.1): Align Data Cube with SDMX 2.1 [Data Cube Vocabulary]

http://www.w3.org/2011/gld/track/issues/37

Raised by: Richard Cyganiak
On product: Data Cube Vocabulary

Data Cube is based on a subset of the SDMX 2.0 information model. The latest version of SDMX is 2.1. The major differences are described in this document:

http://sdmx.org/wp-content/uploads/2011/04/SDMX_2-1_Major_Changes.pdf

Many of the changes are in parts not covered by Data Cube, or are specific to the XML syntaxes and web service / registry functionalities. Some, however, affect the parts covered by Data Cube, and the WG needs to consider whether to update Data Cube to reflect these changes. This would improve compatibility with SDMX data and (assuming the changes happened for good reason) the modelling, at the cost of disrupting legacy data and development time.

The main relevant changes appear to be:

* The attachment level mechanism has changed significantly
* The time period / time format parts of the spec have changed and presumably improved (see Implementation Notes document)
* The mechanism for dimension/attribute roles (isTimeDimension etc.) has been generalized; the role is now picked from a code list, rather than from a hard-coded set
* Measure dimensions are now “more special” than before and defined in terms of a concept scheme rather than code list
* Various differences between cross-sectional and time series datasets have been removed or reduced

Some of these items deal with features that we consciously did not include in Data Cube because they were strange (e.g., the dimension role stuff). We should re-evaluate whether the redesigned feature provides enough value to add it.

Received on Saturday, 13 October 2012 17:45:16 UTC