Development on detect_elements() in 2_process_data.R took place with example definitions that looked for element and visit matches across multiple fields (i.e. defX_list$setup$detect_elements_fields include > 1 field). Ex: ChiefComplaintUpdates, DischargeDiagnosis
In lines 101-108 of 2_process_data.R, multiple detect_elements() output data frames are created (1 dataframe per field), and then bound (via bind_rows()) together by C_BioSense_ID into a main detect_elements data frame with everything to review.
The problem was bind_rows() would result in duplicate column variables C_BSI...1, C_BSI...17, etc --> which had to be taken care of (hence the line 104 causing the error).
Fix
Use reduce(left_join) to eliminate duplicate columns and have this code work even if only 1 field is being used for detect_elements(). Additionally, need to remove creation of TruePositive column within detect_elements() and instead just create it outside of the function (if not then the function would make many TruePositive columns
Background:
Development on
detect_elements()in2_process_data.Rtook place with example definitions that looked for element and visit matches across multiple fields (i.e.defX_list$setup$detect_elements_fieldsinclude > 1 field). Ex:ChiefComplaintUpdates,DischargeDiagnosisIn lines 101-108 of
2_process_data.R, multipledetect_elements()output data frames are created (1 dataframe per field), and then bound (viabind_rows()) together by C_BioSense_ID into a main detect_elements data frame with everything to review.The problem was
bind_rows()would result in duplicate column variables C_BSI...1, C_BSI...17, etc --> which had to be taken care of (hence the line 104 causing the error).Fix
Use
reduce(left_join)to eliminate duplicate columns and have this code work even if only 1 field is being used fordetect_elements(). Additionally, need to remove creation ofTruePositivecolumn withindetect_elements()and instead just create it outside of the function (if not then the function would make manyTruePositivecolumns