More on functions

Useful commands

  • Functions
    • Optional arguments
    • Using ... to accept any arguments
    • Using functions as arguments
  • rbinom()

Functions

Functions carry out a fixed task using the arguments provided:

subtract <- function(value, from) {
  from - value
}
subtract(value = 5, from = 3)
[1] -2
subtract(from = 3, value = 5)
[1] -2

Naming the arguments makes it clearer what you are doing.

Optional arguments

Make an argument optional by providing a default value:

increase <- function(value, by = 1) {
  value + by
}
increase(value = 5, by = 2)
[1] 7
increase(value = 5, by = 1)
[1] 6
increase(value = 5)
[1] 6

Using ... to accept any arguments

Some functions don’t know what their arguments will be:

plot(x, y)
plot(x, y, col="black")
plot(x, y, type="l", lty=2)

The extra arguments get passed on to plot.xy() or some other function, which does then the work.

Using ... to accept any arguments

Some functions don’t know what their arguments will be:

plot(x, y)
plot(x, y, col="black")
plot(x, y, type="l", lty=2)

The extra arguments get passed on to plot.xy() or some other function, which then does the work. It looks a bit like this:

plot <- function(x, y, log = "", ...) {
  xlabel <- deparse1(substitute(x))
  ylabel <- deparse1(substitute(y))
  xy <- xy.coords(x, y, xlabel, ylabel, log)
  plot.xy(xy, ...)
}

Using ... to accept any arguments

Here we do the same in plot_populations():

plot_populations <- function(populations, new.graph=TRUE,
                             ylim=NA, lty=1, col=NA, ...) {
  # -- lots of other code --
  if (new.graph) {
      # When it's a new plot, do labels and legends, etc.
      plot(time, this.pop, ylim = ylim, xlab = "time",
           ylab = "population size", type = "l",
           col = line.cols[index], lty = line.ltys[index], ...)
      # -- more code --
    } else { # Otherwise just draw the lines
      graphics::lines(time, this.pop, col = line.cols[index],
                      lty = line.ltys[index], ...)
    }
}

Using functions as arguments

run_simple <- function(step_function, latest.df, end.time, ...)
{
  population.df <- latest.df
  keep.going <- (latest.df$time < end.time)
  while (keep.going)
  {
    data <- step_function(latest.df, ...)
    latest.df <- data$updated.pop
    population.df <- rbind(population.df, latest.df)
    keep.going <- ((latest.df$time < end.time) &&
                   (!data$end.experiment))
  }
  row.names(population.df) <- NULL
  population.df
}

rbinom()

  • Draw 1 random number
rbinom(1, count, prob)
  • How many infections have occurred if there are 50 animals and there is a probability of infection of 0.1 for each one?
rbinom(1, 50, 0.1)
## [1] 8
rbinom(1, 50, 0.1)
## [1] 5