cct-datascience / ardn-terra-ref Goto Github PK

Task to do

I think we should use names rather than integers for the keys, as you have done with germplasmName;

It is possible / likely that people will prefer these CSV files over the JSON files ... as you have already done the hard work making these tables ready for analysis!

• add studyName and observationName fields; add these to tables that currently have studyDbId and observationDbId.
• events_table doesn't need studyDbId (or study name) since it can be joined via the observations table; its redundancy adds a circular reference and may cause confusion (e.g. people wondering why it is there, or thinking they should join events and studies together on studyDbId).
• please add treatments$factor as treatment or treatmentName to observations because it is required to interpret the data
• update this table relationships figure on google docs and replace one currently in repository

Reason

Consistency, clarity, providing useful information

Result

Hopefully this only updates the tables and doesn't change the downstream steps (e.g. ask vmapper to ignore these extra fields)