4 Data Class

Some R functions require certain kinds of objects as arguments, while other functions can handle many kinds of objects. The latter are called generic functions.

4.1 Generic Functions

Generic functions use different methods according to the class of an object. To view the class of an object, use the class() function:

[1] "data.frame"
mpg <- mtcars$mpg

[1] "numeric"
mod <- lm(mpg ~ wt, data = mtcars)

[1] "lm"

Three commonly used generic functions are print, summary, and plot. Each of these functions has many methods, so the output will vary depending on the class of the object you use.

Try using the print, summary, and plot functions with mtcars, mpg, and mod. What differences do you see?

If you use an object with a class that a function does not handle, R will be happy to give you an error, even though it may be a little cryptic:

Warning in mean.default(x): argument is not numeric or logical: returning NA
Error in var(x): is.atomic(x) is not TRUE
Error in UseMethod("anova"): no applicable method for 'anova' applied to an object of class "data.frame"

On the other hand, you may be surprised at some of the objects that a function does handle! For example, plot will produce a scatterplot matrix when given a data.frame as input.

[1] "data.frame"

4.2 Example - Factor versus Character

Factors and dates are both numeric data, but they are processed in unique ways because of their class attributes. (These specific classes are discussed in more detail in Chapters 8 and 9.)

As an example consider a vector of 25 month names.

cmonth <- sample(month.name, 25, replace = TRUE)
fmonth <- factor(cmonth)

Here cmonth is a character vector, while fmonth has class factor. Compare the output of the summary() and plot() functions when applied to the character vector versus the factor.

   Length     Class      Mode 
       25 character character 
    April    August  December  February   January      July       May  November 
        3         3         3         2         2         4         1         3 
  October September 
        2         2 
Warning in xy.coords(x, y, xlabel, ylabel, log): NAs introduced by coercion
Warning in min(x): no non-missing arguments to min; returning Inf
Warning in max(x): no non-missing arguments to max; returning -Inf
Error in plot.window(...): need finite 'ylim' values


While the printed format of a data object is often a clue as to it’s class, this is not always definitive. In the next example, even though the output of summary(fmonth) and table(fmonth) start with the same data and look the same, the results are stored differently by R. The summary() function gives us a named numeric vector, while table() gives us a table with a numeric vector and a character vector.

tmonth <- table(fmonth)
    April    August  December  February   January      July       May  November 
        3         3         3         2         2         4         1         3 
  October September 
        2         2 
smonth <- summary(fmonth)
    April    August  December  February   January      July       May  November 
        3         3         3         2         2         4         1         3 
  October September 
        2         2 
 'table' int [1:10(1d)] 3 3 3 2 2 4 1 3 2 2
 - attr(*, "dimnames")=List of 1
  ..$ fmonth: chr [1:10] "April" "August" "December" "February" ...
 Named int [1:10] 3 3 3 2 2 4 1 3 2 2
 - attr(*, "names")= chr [1:10] "April" "August" "December" "February" ...

This becomes important when used with the plot() function which will handle these two classes in different ways.



4.3 Example - Date versus Numeric

To take a closer look at dates, we can make a vector of dates from September 1-30, 2020, with seq(). If we coerce the date vector to numeric, we see that R stores dates as the number of days since January 1, 1970.

dates <- seq(from = as.Date("2020-09-01"), 
             to = as.Date("2020-09-30"), 
             by = "days")

ndays <- as.numeric(dates)
 [1] 18506 18507 18508 18509 18510 18511 18512 18513 18514 18515 18516 18517
[13] 18518 18519 18520 18521 18522 18523 18524 18525 18526 18527 18528 18529
[25] 18530 18531 18532 18533 18534 18535

The generic functions we used earlier also handle dates in special ways.

Because dates are numbers, we can calculate summary statistics. There is such a thing as a “mean” date: the average of the dates’ numeric values, reconverted into a date object (and then rounded).

[1] "Date"
 Date[1:30], format: "2020-09-01" "2020-09-02" "2020-09-03" "2020-09-04" "2020-09-05" ...
        Min.      1st Qu.       Median         Mean      3rd Qu.         Max. 
"2020-09-01" "2020-09-08" "2020-09-15" "2020-09-15" "2020-09-22" "2020-09-30" 
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  18506   18513   18521   18521   18528   18535 


The takeaway from these illustrations is that classes play a role in how some functions handle our data, so we need to be aware of the classes of our data objects.