R module to expand a delimited column in a dataframe into multiple binary columnar output


Imagine you have a data frame that contains a column which has a variety of delimited values in a single column. Note, this may take a while on large data sets, and is not yet optimized for performance. E.g. a 90k row data set where the column splits into 13 sub-columns may take up to 1-2 minutes to run as the function generates each new sub-column in the data frame.

Example Case

Example: The column lists “My Favorite Programming Languages” in a column titled “lang”. The individuals who were surveyed could have entered one or more programming languages as their favorite. An example of survey results could look like the following.

       User 1: Python, Java
       User 2: Java, C
       User 3: Python

       And in the data set, we would see something like:
        | User   | lang        |
        | user_1 | Python;Java |
        | user_2 | Java;C      |
        | user_3 | Python      | 
       While efficient for data storage, it is not easy to do any analysis
       on this data set without splitting out the exact values. One
       issue found with most modern libraries is that they are expecting
       your values to be in a specific order when they are split. 
       This method does not make that assumption.
       dfexpand's expand_column() function will extract the unique
       values found for all users surveyed from this column, then it will
       iterate and expand the column from "lang" into 3 new distinct columns.
       New columns:
         "lang_Python", "lang_Java", and "lang_C"

       Further, it will populate each new column with a 0 or 1
       as appropriately matched to each row entry.

       For the above example, you would then expect to see in the expanded dataframe:

        | User   | lang_Python |  lang_Java  |  lang_C |
        | user_1 |       1     |       1     |     0   |
        | user_2 |       0     |       1     |     1   |
        | user_3 |       1     |       0     |     0   |

      The method does not drop the original 'lang' column but it has not been
      shown here for simplification.

Example R code

    myDelimiter = ";"

    # Create some fake data with duplicates
    rows = c(
        c("a;b"), c("a;b;c"), c("b;c"), c("d"), c("d")

    # Add to a dataframe
    df = data.frame(rows)

    colnames(df) <- c("myvar")
    # The default behavior is to trim extra whitespace from the extracted values, 
    # but not to alter or change the case of the values. So 'Alpha' is distinct from 'alpha'
    # but ' beta ' is the same as 'beta'. You can override this behavior with
    # the trim and ignore case flags.
    expanded_df = expand_column(df, "myvar", myDelimiter)