Open fiendish opened 2 years ago
I hypothesize that this should be totally possible because we know how the different standard concepts are related to each other. One way might be to just throw all the dataframes into the graph structure used by the validator and then extract the result into a new set of dataframes. Alternatively, the right sequence of table joins on the right identifier columns should be able to get all of the attributes in the same tables as their IDs (but watch out for the traps described in https://github.com/kids-first/kf-lib-data-ingest/blob/master/docs/archive/design/merge_algorithm/DataFrame_Design_Complications_Motivating_The_Relationship_Graph.ipynb)
This would be amazingggg if it gets implemented!
I hypothesize that this should be totally possible because we know how the different standard concepts are related to each other. One way might be to just throw all the dataframes into the graph structure used by the validator and then extract the result into a new set of dataframes. Alternatively, the right sequence of table joins on the right identifier columns should be able to get all of the attributes in the same tables as their IDs (but watch out for the traps described in https://github.com/kids-first/kf-lib-data-ingest/blob/master/docs/archive/design/merge_algorithm/DataFrame_Design_Complications_Motivating_The_Relationship_Graph.ipynb)