presently the database server looks like this:
Ideally we move towards :
databases:
dev (developmental) database
- work on creating new tables / columns / relationships
prod (production) database
- static and we can dump the database and have a commit ID and even a DOI
- copy of dev for a given timestamp
schemas:
raw
raw data, before cleaning / processing or adding any derived columnspublic
final tables for use with APIs, etc; ideally all modifications and QA/QC are generated from reproducible scripts (in workflows)
presently the database server looks like this:
Ideally we move towards :
databases:
dev(developmental) databaseprod(production) databaseschemas:
rawraw data, before cleaning / processing or adding any derived columns
publicfinal tables for use with APIs, etc; ideally all modifications and QA/QC are generated from reproducible scripts (in workflows)