Allow arbitrary (non-component) columns in cube-pipeline
Robsteranium opened this issue · comments
Our interpretation of kubus luchtemissies in the examples doesn't reproduce the original faithfully. The original data includes several statements for every observation where the property isn't a component in the cubes DSD, for example rdfs:label.
We might also want to include arbitrary columns in the cube-pipeline so that they could be used in URI templates - e.g. SHAs in the observation-uri.
One candidate for configuring non-component columns would be to leave a blank value in the "component_attachment" column. The we could just leave the property out of the DSD (and not create a component spec etc). The problem with this is that that currently signifies a value column. Indeed currently, if you do this in the above example, you get:
Columns Label represent observation values. Multi-measure cubes should define measure values in the corresponding measure columns.
We should investigate to see if it's possible to support this (e.g. things like table2qb.pipelines.cube/suppress-value-column would become suppress-non-component-column).
NB: In the mean time, arbitrary properties could be added to observations by simply making them attribute properties.