Add correspondence table(s)

[cmutel]

Bonsai policy is to store data on its "native" scale, which means we need correspondence from one set to another. Traditionally, we would have 1 to 1 and 1 to N, but maybe it also makes sense to have approximate 1 to 1 matches, and to distinguish 1 to 1 that is official versus matches made by Bonsai team members.

@mfastudillo You know this as well as anyone, want to suggest something? We would need a table name and the names and data types of the columns, plus values to be added whenever we want to limit choices to a select few (enumerated types).

[mfastudillo]

sorry, I am a bit lost and I may be completely missing the point.

So for example, if we put the wri global power plant database on our relational bonsai database we will have among others a table of activities. To pick an example, one of the activities will be "electricity generation" coming from a plant whose "primary_fuel" is Gas... then we'll need a correspondence table claryfing that what they call Gas in the Bonsai ontology is xxx (as far as I know we have not adopted an otology of power plants yet) ?

[cmutel]

No, you have the point exactly.

Our policy is to store raw data on its native scale (decided at 2019 hackathon). To be useful, that data musts be aligned with a common core nomenclature (i.e. naming system, not ontology; that is things like products versus agents versus activities). We don't have a fixed common core nomenclature but something like HS tariff product codes, ISIC industry code, ISO country/region codes seems reasonable.

The correspondence table would be something like:

Source name	Dest name	Relationship
Electricity (gas)	Electricity, from gas	exact
Electricity (gas)	Electricity	narrower

This issues is about how we structure this table - column names, number of columns, what terms to use, etc.