-
Notifications
You must be signed in to change notification settings - Fork 62
Description
The current function FeatureExtraction::getDbCovariateData forces the download of the covariates. The default covariate function FeatureExtraction:::getDbDefaultCovariateData has inputs such as targetCovariateTable that let the user skip the download of the covariates and instead inserts them into the specified database tables.
The cohorts as features covariate function does not let you specify a database and table to save results to.
It would be nice if all the covariates get functions such as FeatureExtraction:::getDbCohortBasedCovariatesData had these inputs:
- targetDatabaseSchema
- targetCovariateTable
- targetCovariateRefTable
- targetAnalysisRefTable
and when specified inserted the results to a database table rather than downloaded. Then there could be a new function such as FeatureExtraction::insertDbCovariateData that takes the same inputs as FeatureExtraction::getDbCovariateData but also has:
- targetDatabaseSchema
- targetCovariateTable
- targetCovariateRefTable
- targetAnalysisRefTable
and instead of downloading the covariates inserts them all into the specified database and set of tables.
This would help any code that does further manipulation of the features as that can be done in SQL rather than R.