Skip to content

Holistic review of attachments for data attributes and for DSD-extending metadata attributes #58

@dosse

Description

@dosse

Reference documents:

Problem statement:

  • Unique (ref. metadata) attribute values per DSD respond to a business need
  • Also dataflow-level (ref. metadata) attribute (values) (especially those with empty keys (“..~”)) respond to a business need
  • But attaching (ref. metadata) attribute (values) to a dataflow (as currently defined in the DSD) creates conflicts

3 alternative solutions proposed:

  • Full context preservation - approach
    • All 3, DF, DSD and PA-specific (ref. metadata) attribute values are allowed for any attachment (full, partial, empty key)
    • The current “DF-level” attribute attachment becomes “dimension-independent”
  • Partial context preservation - approach - Due to inconsistencies, this solutions was discarded by the TWG (Jan 2026 meeting)
    • All 3, DF, DSD and PA-specific (ref. metadata) attribute values are allowed for attachments without dimensions, e.g., “..~”, thus all partial and full attachments remain unique per DSD
    • The current “DF-level” attribute attachment becomes “dimension-independent”
  • Only unique values per DSD - approach
    • Only DSD-specific (ref. metadata) attribute values are allowed whatever the attachment (full, partial or empty, e.g., “..~”), the corresponding values are unique per DSD
    • The current “DF-level” attribute attachment is promoted to DSD-level
    • DF-level (ref. metadata) attributes (defined with DSD/MSD) are not supported
    • DF-specific ref. metadata can be provided through MSD/MDF/Report

There might be contradicting needs thus separate mechanisms might be necessary for context-specific values and for simple views on sub-cubes.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions