Dealing with attachmentLevel Series

[csarven]

If @attachmentLevel='Series' for a Component, does it correspond to <qb:componentAttachment rdf:resource="{$qb}Dimension"/>?

[cygri]

Do you have an example SDMX file for this? (I believe the answer is yes, but would like to double-check)

[csarven]

Examples in repo:
data/BIS.WEBSTATS_CIBL_UR_DATAFLOW-1351443106306.xml
data/ECB.KeyFamily.xml
data/IMF.BOPDSD.xml
data/OECD.HEALTH_STAT.xml
data/SDMX.StructureSample.xml

See also cube.ttl:

<http://purl.org/linked-data/cube#componentAttachment> <http://www.w3.org/2000/01/rdf-schema#range <http://www.w3.org/2000/01/rdf-schema#Class> .
<http://purl.org/linked-data/cube#componentAttachment> <http://www.w3.org/2000/01/rdf-schema#comment> "Indicates the level at which the component property should be attached, this might an qb:DataSet, qb:Slice or qb:Observation, or a qb:MeasureProperty."@en .

The range is flexible but the comment is a bit restrictive.

If I'm not mistaken, QB is missing an attachment level at the Series level. It would be similar to qb:Slice level but definitely at a higher level than qb:Observation.

[cygri]

Reading the relevant parts of QB and SDMX again, there are two ways of handling this.

First, you can always “flatten” the dataset, and attach everything to the observations. The attachment level of the component would be set to qb:Observation. The triple with the value has to be repeated on every applicable observation.

The other option is to define a qb:SliceKey that covers the time dimension, and then set the attachment level to qb:Slice. That's because slices in data cube are a mechanism that unifies groups and series. You would then actually have to define qb:Slice instances for each series, and attach the attribute triple there.