Control-flow.Rmd

# Control flow

```{r, include = FALSE}
source("common.R")
```

## Introduction

There are two primary tools of control flow: choices and loops. Choices, like `if` statements and `switch()` calls, allow you to run different code depending on the input. Loops, like `for` or `while`, allow you to repeatedly run code, typically with changing options. I expect that you're already familiar with the basics of these functions so I'll briefly cover some technical details and introduce you to some useful, but lesser known, features.

The condition system (messages, warnings, and errors) also provides "non-local" control flow. You'll learn about them in Chapter \@ref(conditions). 

### Quiz {-}

Want to skip this chapter? Go for it, if you can answer the questions below. Find the answers at the end of the chapter in Section \@ref(control-flow-answers).

*   What is the different between `if` and `ifelse()`?

*   In the following code, what will the value of `y` be if `x` is `TRUE`?
    What if `x` is `FALSE`? What if `x` is `NA`?
  
    ```{r, eval = FALSE}
    y <- if (x) 3
    ```

*   What does `switch("x", x = , y = 2, z = 3)` return?

### Outline {-}

* Section \@ref(choices) dives into the details of `if`, then discusses
  the close relatives `ifelse()` and `switch()`.
  
* Section \@ref(loops) starts off by reminding you of the basic structure
  of the for loop in R, discusses some common pitfalls, and then talks
  about the related `while` and `repeat` statements.

## Choices
\indexc{if}

The basic form of R's if statement is as follows:

```{r, eval = FALSE}
if (condition) true_action
if (condition) true_action else false_action
```

If `condition` is `TRUE`, `true_action` will be evaluated; if `condition` is `FALSE`, the optional `else` statement will be evaluated. 

Typically the actions are compound statements contained within `{`:

```{r}
grade <- function(x) {
  if (x > 90) {
    "A"
  } else if (x > 80) {
    "B"
  } else if (x > 50) {
    "C"
  } else {
    "F"
  }
}
```

`if` returns a value so that you can assign the results:

```{r}
x1 <- if (TRUE) 1 else 2
x2 <- if (FALSE) 1 else 2

c(x1, x2)
```

(I recommend using assigning the results of an entire `if` statement only when the whole thing fits one line; otherwise it tends to be hard to read.)

When you use the single argument form without an else statement, `if` invisibly (Section \@ref(invisible)) returns `NULL` if the condtion is `FALSE`. Since functions like `c()` and `paste()` drop `NULL` inputs, this allows for a compact expression of certain idioms:

```{r}
greet <- function(name, birthday = FALSE) {
  paste0(
    "Hi ", name,
    if (birthday) " and HAPPY BIRTHDAY"
  )
}
greet("Maria", FALSE)
greet("Jaime", TRUE)
```

### Invalid inputs

The `condition` should evaluate to a single `TRUE` or `FALSE`. Most other inputs generate an error:

```{r, error = TRUE}
if ("x") 1
if (logical()) 1
if (NA) 1
```

The exception is a logical vector of length greater than 1, which only generates a warning:

```{r, include = FALSE}
Sys.setenv("_R_CHECK_LENGTH_1_CONDITION_" = "false")
```

```{r}
if (c(TRUE, FALSE)) 1
```

In R 3.5.0 and greater, thanks to [Henrik Bengtsson](https://github.com/HenrikBengtsson/Wishlist-for-R/issues/38), you can turn this an error by setting an environment variable:

```{r, error = TRUE}
Sys.setenv("_R_CHECK_LENGTH_1_CONDITION_" = "true")
if (c(TRUE, FALSE)) 1
```

I think it's good practice to set this environment variable, as it reveals a clear mistake that you might miss if only shown as a warning.

### Vectorised if
\indexc{ifelse()}

Given that `if` only works with a single `TRUE` or `FALSE`, you might wonder what to do if you have a vector of logical values. Handling vectors of values is the job of `ifelse()`: a vectorised function with `test`, `yes`, and `no` vectors (that will be recycled to the same length):

```{r}
x <- 1:10
ifelse(x %% 5 == 0, "XXX", as.character(x))

ifelse(x %% 2 == 0, "even", "odd")
```

Missing values will be propagated into the output.

I recommend only using `ifelse()` when the `yes` and `no` vectors are the same type, as it is otherwise hard to predict the output type. See about <https://vctrs.r-lib.org/articles/stability.html#ifelse> for additional discussion.

Another vectorised equivalent is the more general `dplyr::case_when()`. It uses a special syntax to allow any number of condition-vector pairs:

```{r}
dplyr::case_when(
  x %% 35 == 0 ~ "fizz buzz",
  x %% 5 == 0 ~ "fizz",
  x %% 7 == 0 ~ "buzz",
  is.na(x) ~ "???",
  TRUE ~ as.character(x)
)
```

### `switch()` statement {#switch}
\indexc{switch()}

Closely related to `if` is the `switch()`-statement. It's a compact, special purpose equivalent that lets you replace code like:

```{r}
x_option <- function(x) {
  if (x == "a") {
    "option 1"
  } else if (x == "b") {
    "option 2" 
  } else if (x == "c") {
    "option 3"
  } else {
    stop("Invalid `x` value")
  }
}
```

with the more succinct:

```{r}
x_option <- function(x) {
  switch(x,
    a = "option 1",
    b = "option 2",
    c = "option 3",
    stop("Invalid `x` value")
  )
}
```

The last component of a `switch()` should always throw an error, otherwise unmatched inputs will invisibly return `NULL`:

```{r}
(switch("c", a = 1, b = 2))
```

If multiple inputs have the same output, you can leave the right hand side of `=` empty and the input will "fall through" to the next value. This mimics the behaviour of C's `switch` statement:

```{r}
legs <- function(x) {
  switch(x,
    cow = ,
    horse = ,
    dog = 4,
    human = ,
    chicken = 2,
    plant = 0,
    stop("Unknown input")
  )
}
legs("cow")
legs("dog")
```

It is also possible to use `switch()` with numeric `x`, but is harder to read, and has undesirable failure modes if `x` is a not a whole number. I recommend using `switch()` only with character inputs.

### Exercises

1.  What type of vector does each of the following calls to `ifelse()`
    return?

    ```{r, eval = FALSE}
    ifelse(TRUE, 1, "no")
    ifelse(FALSE, 1, "no")
    ifelse(NA, 1, "no")
    ```

    Read the documentation and write down the rules in your own words.

1.  Why does the following code work?

    ```{r}
    x <- 1:10
    if (length(x)) "not empty" else "empty"
      
    x <- numeric()
    if (length(x)) "not empty" else "empty"
    ```

## Loops
\index{loops}
\index{loops!for@\texttt{for}}
\indexc{for}

For loops are used to iterate over items in a vector. They have the following basic form:

```{r, eval = FALSE}
for (item in vector) perform_action
```

`perform_action` is called once for each item in `vector`, rebinding the value of `item` each time.

```{r}
for (i in 1:3) {
  print(i)
}
```

(When iterating over a vector of indices, it's conventional to use very short variable names like `i`, `j`, or `k`.)

NB: `for` assigns the `item` in the current environment, so that an existing variable with the same name will be overwritten:

```{r}
i <- 100
for (i in 1:3) {}
i
```

\indexc{next}
\indexc{break}
There are two ways to terminate a for loop early:

* `next` terminates the current iteration and advances to the next.
* `break` exits the entire for loop.

```{r}
for (i in 1:10) {
  if (i < 3) 
    next

  print(i)
  
  if (i >= 5)
    break
}
```

### Common pitfalls
\index{loops!common pitfalls}

There are three common problems that you might encounter when using `for`. First, if you're generating data, make sure to preallocate the output. Otherwise the loop will be very slow; see Sections \@ref(memory-profiling) and \@ref(avoid-copies) for more details. The `vector()` function is helpful here.

```{r}
means <- c(1, 50, 20)
out <- vector("list", length(means))
for (i in 1:length(means)) {
  out[[i]] <- rnorm(10, means[[i]])
}
```

Next, beware iterating over `1:length(x)`. This will fail in unhelpful ways if `x` has length 0:

```{r, error = TRUE}
means <- c()
out <- vector("list", length(means))
for (i in 1:length(means)) {
  out[[i]] <- rnorm(10, means[[i]])
}
```

That's because `:` works with both positive and increasing sequences:

```{r}
1:length(means)
```

Instead switch to `seq_along(x)` which always returns a value the same length as `x`:

```{r}
seq_along(means)

out <- vector("list", length(means))
for (i in seq_along(means)) {
  out[[i]] <- rnorm(10, means[[i]])
}
```

Finally, you might encounter problems when iterating over S3 vectors, as loops typically strip the attributes:

```{r}
xs <- as.Date(c("2020-01-01", "2010-01-01"))
for (x in xs) {
  print(x)
}
```

Work around this by calling `[[` yourself:

```{r}
for (i in seq_along(xs)) {
  print(xs[[i]])
}
```

### Related tools {#for-family}
\indexc{while}
\indexc{repeat}

`for` loops are useful if you know in advance the set of values that you want to iterate over. There are two related tools with more flexible specifications:

* `while(condition) action`: performs `action` while `condition` is `TRUE`.

* `repeat(action)`: repeats action forever (i.e. until it encounters `break`).

R does not have an equivalent to the `do {action} while (condition)` syntax found in other languages.

You can rewrite any `for` loop to use `while` instead, and you can rewrite any `while` loop to use `repeat`, but the converses are not true. That means `while` is more flexible than `for`, and `repeat` is more flexible than `while`. It's good practice to use the least-flexible solution to a problem, which means that you should use `for` where possible.

Generally speaking you should not need to use for loops for data analysis tasks, as `map()`/`apply()` functions provide an even less flexible solution to most problems. You'll learn more in Chapter \@ref(functionals).

### Exercises

1.  Why does this code succeed without errors or warnings? 
    
    ```{r, results = FALSE}
    x <- numeric()
    out <- vector("list", length(x))
    for (i in 1:length(x)) {
      out[i] <- x[i] ^ 2
    }
    out
    ```

1.  What does the following code tell you about when the vector being 
    iterated over is evaluated?

    ```{r}
    xs <- c(1, 2, 3)
    for (x in xs) {
      xs <- c(xs, x * 2)
    }
    xs
    ```

1.  What does the following code tell you about when the index is updated?

    ```{r}
    for (i in 1:3) {
      i <- i * 2
      print(i) 
    }
    ```

## Answers {#control-flow-answers}

* `if` works with scalars; `ifelse()` works with vectors.

* When `x` is `TRUE`, `y` will be `3`, when `FALSE` `y` will be `NULL`,
  when `NA` the if statement will throw an error.

* This `switch()` statement makes use of fall-through so it will return 2.
  See details in Section \@ref(switch).