Skip to content

Feature Request: Detect Columns that Map to Each Other #291

@billdenney

Description

@billdenney

I often receive data with columns that contain either identical or effectively identical columns.

As an example, I may have a column indicating a human-readable medical laboratory test (like "Cholesterol, LDL"), and another column indicating a more computer-friendly standardized test number (called a LOINC code that may be 12345). It would be helpful to me to have a reporter function that indicates the columns that all map to the same values. The interface could look like:

find_mapping_cols(x)

where it would return a list with elements that are character vectors of columns that map to each other:

foo <- data.frame(
  Lab_Test_Long=c("Cholesterol, LDL", "Cholesterol, LDL", "Glucose"),
  Lab_Test_Short=c("CLDL", "CLDL", "GLUC"),
  LOINC=c(12345, 12345, 54321),
  Person=c("Sam", "Bill", "Sam"),
  stringsAsFactors=FALSE
)
find_mapping_cols(foo)

That would return a list that looks like:

list(c("Lab_Test_Long", "Lab_Test_Short", "LOINC"), "Person")

It's up to the user what they would want to do with that list (because which column of the set is most useful to keep is not obvious).

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions