In the below lines, the age group covariate name is formed by assuming that an int will be returned when 2 divisible ints are divided. In Spark SQL, unfortunately, this is not the case; rather, a double is produced, which then throws off the covariate name.
In the below lines, the age group covariate name is formed by assuming that an int will be returned when 2 divisible ints are divided. In Spark SQL, unfortunately, this is not the case; rather, a double is produced, which then throws off the covariate name.
https://github.com/OHDSI/FeatureExtraction/blob/master/inst/sql/sql_server/DemographicsAgeGroup.sql#L37-L39
Suggesting to explicitly cast the division to an integer before casting to varchar.