Closed dosse closed 3 years ago
@jgager In today's TWG meeting, a question was raised for you, if you remember specific reasons why this attribute was added and that would need to be considered in the decision for this ticket. Thanks in advance.
@sdmx3mdt Could you please notify @jgager of this question, if required?
@dosse I don't recall anything specific nor do I find any notes relating specifically to it. If I had to hazard a guess, I would say that the intention was to be specific about which constraints were followed, if any.
I agree that it is redundant with the structure reference, since that indicates if that data is reported against a structure, flow, or provision agreement.
Thanks for the quick response, @jgager!
@sdmx3mdt If there is no other follow-up reaction from the TWG, then we could consider the removal of the attribute as approved?
@dosse thanks - TWG approved the change in principle so with no technical restrictions we can simply go ahead and apply it. @jgager I'll create a ticket on the SDMX-ML working repo for this.
Schemas updated to remove redundant dataScope attribute from the data message.
It seems that the dataSet property dataScope only exists in the 2.1 xsd schema but not in the SDMX Information Model. Since it doesn't bring any additional information on top of the information in the message header references (dsd, df or pa) and worst, it potentially clashes with it, it should be removed.
The current DataScopeType is defined as:
Note that it's "DataStructure" value actually allows a message content that violates the SDMX information model. If a constraint has been defined on the data (at whatever level DSD, DF or PA) then allowing a message to carry data not respecting that constraint is a security flaw.