r4ds/strings.Rmd

# Strings

```{r, results = "asis", echo = FALSE}
status("restructuring")
```

## Introduction

So far, you've used a bunch of strings without learning much about the details.
Now it's time to dive into them, learning what makes strings tick, and mastering some of the powerful string manipulation tool you have at your disposal.

We'll begin with the details of creating strings and character vectors.
You'll then learn a grab bag of handy string functions before we dive into creating strings from data.
Next, we'll discuss the basics of regular expressions, a powerful tool for describing patterns in strings, then use those tools to extract data from strings.
The chapter finishes up with a brief discussion where English language expectations might steer you wrong when working with text from other languages.

This chapter is paired with two other chapters.
Regular expression are a big topic, so we'll come back to them again in Chapter \@ref(regular-expressions).
We'll come back to strings again in Chapter \@ref(programming-with-strings) where we'll think about them about more from a programming perspective than a data analysis perspective.

### Prerequisites

In this chapter, we'll use functions from the stringr package.
The equivalent functionality is available in base R (through functions like `grepl()`, `gsub()`, and `regmatches()`) but we think you'll find stringr easier to use because it's been carefully designed to be as consistent as possible.
We'll also work with the babynames dataset since it provides some fun data to apply string manipulation to.

```{r setup, message = FALSE}
library(tidyverse)
library(babynames)
```

You can easily tell when you're using a stringr function because all stringr functions start with `str_`.
This is particularly useful if you use RStudio, because typing `str_` will trigger autocomplete, allowing you jog your memory of which functions are available.

```{r, echo = FALSE}
knitr::include_graphics("screenshots/stringr-autocomplete.png")
```

## Creating a string

We've created strings in passing earlier in the book, but didn't discuss the details.
First, there are two basic ways to create a string: using either single quotes (`'`) or double quotes (`"`).
Unlike other languages, there is no difference in behaviour, but the [tidyverse style guide](https://style.tidyverse.org/syntax.html#character-vectors) recommends using `"`, unless the string contains multiple `"`.

```{r}
string1 <- "This is a string"
string2 <- 'If I want to include a "quote" inside a string, I use single quotes'
```

If you forget to close a quote, you'll see `+`, the continuation character:

    > "This is a string without a closing quote
    + 
    + 
    + HELP I'M STUCK

If this happen to you and you can't figure out which quote you need to close, press Escape to cancel, and try again.

### Escapes

To include a literal single or double quote in a string you can use `\` to "escape" it:

```{r}
double_quote <- "\"" # or '"'
single_quote <- '\'' # or "'"
```

So if you want to include a literal backslash in your string, you'll need to double it up: `"\\"`:

```{r}
backslash <- "\\"
```

Beware that the printed representation of a string is not the same as string itself, because the printed representation shows the escapes.
To see the raw contents of the string, use `str_view()` [^strings-1]:

[^strings-1]: You can also use the base R function `writeLines()`

```{r}
x <- c(single_quote, double_quote, backslash)
x
str_view(x)
```

### Raw strings

Creating a string with multiple quotes or backslashes gets confusing quickly.
To illustrate the problem, lets create a string that contains the contents of the chunk where I define the `double_quote` and `single_quote` variables:

```{r}
tricky <- "double_quote <- \"\\\"\" # or '\"'
single_quote <- '\\'' # or \"'\""
str_view(tricky)
```

That's a lot of backslashes!
(I like the evocative name for this problem: [leaning toothpick syndome](https://en.wikipedia.org/wiki/Leaning_toothpick_syndrome))

To eliminate the escaping you can instead use a **raw string**[^strings-2]:

[^strings-2]: Available in R 4.0.0 and above.

```{r}
tricky <- r"(double_quote <- "\"" # or '"'
single_quote <- '\'' # or "'")"
str_view(tricky)
```

A raw string usually starts with `r"(` and finishes with `)"`.
But if your string contains `)"` you can instead use `r"[]"` or `r"{}"`, and if that's still not enough, you can insert any number of dashes to make the opening and closing pairs unique, e.g. `` `r"--()--" ``, `` `r"---()---" ``, etc. Raw strings are flexible enough to handle any text.

### Other special characters

As well as `\"`, `\'`, and `\\` there are a handful of other special characters that may come in handy. The most common are `\n`, newline, and `\t`, tab, but you can see the complete list in `?'"'`.
You'll also sometimes see strings containing Unicode escapes that start with `\u` or `\U`.
This is a way of writing non-English characters that works on all systems:

```{r}
x <- c("\u00b5", "\U0001f604")
x
str_view(x)
```

### Vectors

You can combine multiple strings into a character vector by using `c()`:

```{r}
x <- c("first string", "second string", "third string")
x
```

Technically, a string is a length-1 character vector, but this doesn't have much bearing on your data analysis live.

If needed, you can create a length zero character vector with `character()`.
This is not generally very useful, but because it's the shortest possible vector, it can sometimes be useful for determining the general pattern of a function by feeding it an extreme.

Now that you've learned the basics of creating strings by "hand", we'll go into the details of creating strings from other strings, starting with a grab bag of small, but useful functions.

### Exercises

## Handy functions

### Length

It's often natural to think about the letters that make up an individual string.
`str_length()` tells you the length of a string:

```{r}
str_length(c("a", "R for data science", NA))
```

You could use this with `count()` to find the distribution of lengths of US babynames, and then with `filter()` to look at the longest names[^strings-3]:

[^strings-3]: Looking at these entries, I'd say the babynames data removes spaces or hyphens from names and truncates after 15 letters.

```{r}
babynames %>%
  count(length = str_length(name), wt = n)

babynames %>% 
  filter(str_length(name) == 15) %>% 
  count(name, wt = n, sort = TRUE)
```

### Long strings

Sometimes the reason you care about the length of a string is because you're trying to fit it into a label on a plot or in a table.
stringr provides two useful tools for cases where your string is too long:

-   `str_trunc(x, 20)` ensures that no string is longer than 20 characters, replacing any thing too long with `…`.

-   `str_wrap(x, 20)` wraps a string introducing new lines so that each line is at most 20 characters (it doesn't hyphenate, however, so any word longer than 20 characters will make a longer time)

```{r}
x <- "Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat."

str_trunc(x, 30)
str_view(str_wrap(x, 30))
```

### Subsetting

You can extract parts of a string using `str_sub(string, start, end)`.
The `start` and `end` arguments are inclusive, so the length of the returned string will be `end - start + 1`:

```{r}
x <- c("Apple", "Banana", "Pear")
str_sub(x, 1, 3)
```

You can use negative values to count back from the end of the string: -1 is the last character, -2 is the second to last character, etc.

```{r}
str_sub(x, -3, -1)
```

Note that `str_sub()` won't fail if the string is too short: it will just return as much as possible:

```{r}
str_sub("a", 1, 5)
```

We could use `str_sub()` with `mutate()` to find the first and last letter of each name:

```{r}
babynames %>% 
  mutate(
    first = str_sub(name, 1, 1),
    last = str_sub(name, -1, -1)
  )
```

### Exercises

1.  Use `str_length()` and `str_sub()` to extract the middle letter from each baby name. What will you do if the string has an even number of characters?
2.  Are there any major trends in the length of babynames over time? What about the popularity of first and last letters?

## Creating strings from data

There are two ways in which you might want to combine strings.
You might have a few character vectors which you want to combine together creating a new vector.
Or you might have a single vector that you want to collapse down into a single string.

### `str_c()`

Use `str_c()`[^strings-4] to join together multiple character vectors into a single vector:

[^strings-4]: `str_c()` is very similar to the base `paste0()`.
    There are two main reasons I recommend: it obeys the usual rules for handling `NA` and it uses the tidyverse recycling rules.

```{r}
str_c("x", "y")
str_c("x", "y", "z")
```

`str_c()` is designed to be used with `mutate()` so it obeys the usual tidyverse recycling and missing value rules:

```{r}
df <- tibble(name = c("Timothy", "Dewey", "Mable", NA))
df %>% mutate(greeting = str_c("Hi ", name, "!"))
```

If you want missing values to display in some other way, use `coalesce()` either inside or outside of `str_c()`:

```{r}
df %>% mutate(
  greeting1 = str_c("Hi ", coalesce(name, "you"), "!"),
  greeting2 = coalesce(str_c("Hi ", name, "!"), "Hi!")
)
```

### `str_glue()`

One of the downsides of `str_c()` is that you have to constantly open and close the string in order to include variables.
An alternative approach is provided by the glue package with `str_glue()`[^strings-5] .
Glue works a little differently to `str_c()`: you give it a single string that uses `{}` to indicate where variables should be evaluated:

[^strings-5]: If you're not using stringr, you can also access it directly with `glue::glue().`

```{r}
df %>% mutate(greeting = str_glue("Hi {name}!"))
```

You can use any valid R code inside of `{}`, but it's a good idea to pull complex calculations out into their own variables so you can more easily check your work.

Currently, `str_glue()` is slightly inconsistent `str_c()` but we'll hopefully fix that by the time the book is printed: <https://github.com/tidyverse/glue/issues/246>

### `str_flatten()`

`str_c()` and `glue()` are work well with `mutate()` because the output is the same length as the input.
What if you want a function that works well with `summarise()`, a function who's output is always length 1, regardless of the length of the input?

That's the job of `str_flatten()`:[^strings-6] it takes a character vector and always returns a single string:

[^strings-6]: The base R equivalent is `paste()` with the `collapse` argument set.

```{r}
str_flatten(c("x", "y", "z"))
str_flatten(c("x", "y", "z"), ", ")
str_flatten(c("x", "y", "z"), ", ", last = ", and ")
```

Which makes it work well with `summarise()`:

```{r}
df <- tribble(
  ~ name, ~ fruit,
  "Carmen", "banana",
  "Carmen", "apple",
  "Marvin", "nectarine",
  "Terence", "cantaloupe",
  "Terence", "papaya",
  "Terence", "madarine"
)
df %>%
  group_by(name) %>% 
  summarise(fruits = str_flatten(fruit, ", "))
```

### Exercises

1.  Compare and contrast the results of `paste0()` with `str_c()` for the following inputs:

    ```{r, eval = FALSE}
    str_c("hi ", NA)
    str_c("hi ", character())
    str_c(letters[1:2], letters[1:3])
    ```

2.  What does `str_flatten()` return if you give it a length 0 character vector?

## Working with patterns

Before we can discuss the opposite problem of extracting data out of strings, we need to take a quick digression to talk about **regular expressions**.
Regular expressions are a very concise language for describing patterns in strings.

### Detect matches

To determine if a character vector matches a pattern, use `str_detect()`.
It returns a logical vector the same length as the input:

```{r}
x <- c("apple", "banana", "pear")
str_detect(x, "e")
```

This makes it a logical pairing with `filter()`.
The following example returns all names that contain a lower-case "x":

```{r}
babynames %>% filter(str_detect(name, "x"))
```

Remember that when you use a logical vector in a numeric context, `FALSE` becomes 0 and `TRUE` becomes 1.
That means you can use `summarise()` with `sum()` or `mean()` and `str_detect()` if you want to answer questions about the prevalence of patterns.
For example, the following snippet, gives the proportion of names containing an "x" by year:

```{r, fig.alt = "A timeseries showing the proportion of baby names that contain the letter x. The proportion declines gradually from 8 per 1000 in 1880 to 4 per 1000 in 1980, then increases rapidly to 16 per 1000 in 2019."}
babynames %>% 
  group_by(year) %>% 
  summarise(prop_x = mean(str_detect(name, "x"))) %>% 
  ggplot(aes(year, prop_x)) + 
  geom_line()
```

(Note that this gives us the proportion of names that contain an x; if you wanted the proportion of babies given a name containing an x, you'd need to perform a weighted mean).

### Introduction to regular expressions

To understand what's going on, we need to discuss what the second argument to `str_detect()` really is.
It looks like a simple string, but it's pattern actually a much richer tool called a **regular expression**.
A regular expression uses special characters to match string patterns.
For example, `.` will match any character, so `"a."` will match any string that contains an a followed by another character:

```{r}
str_detect(c("a", "ab", "ae", "bd", "ea", "eab"), "a.")
```

`str_view()` shows you regular expressions to help understand what's happening:

```{r}
str_view(c("a", "ab", "ae", "bd", "ea", "eab"), "a.")
```

Regular expressions are a powerful and flexible language which we'll come back to in Chapter \@ref(regular-expressions).
Here we'll use only the most important components of the syntax as you learn the other stringr tools for working with patterns.

There are three useful **quantifiers** that can be applied to other pattern: `?` makes a pattern option (i.e. it matches 0 or 1 times), `+` lets a pattern repeat (ie. it matches at least once), and `*` lets a pattern be optional or repeat (i.e. it matches any number of times, including 0).

-   `ab?` match an "a", optionally followed by a b

-   `ab+` matches an "a", followed by at least one b

-   `ab*` matches an "a", followed by any number of bs

There are various alternatives to `.` that match a restricted set of characters.
One useful operator is the **character class:** `[abcd]` match "a", "b", "c", or "d"; `[^abcd]` matches anything **except** "a", "b", "c", or "d".

You can opt-out of the regular expression rules by using `fixed`:

```{r}
str_view(c("", "a", "."), fixed("."))
```

Note that both fixed strings and regular expressions are case sensitive by default.
You can opt out by setting `ignore_case = TRUE`.

```{r}
str_view_all("x  X  xy", "X")
str_view_all("x  X  xy", fixed("X", ignore_case = TRUE))
str_view_all("x  X  xy", regex(".Y", ignore_case = TRUE))
```

We'll come back to case later, because it's not trivial for many languages.

### Count matches

A variation on `str_detect()` is `str_count()`: rather than a simple yes or no, it tells you how many matches there are in a string:

```{r}
str_count(x, "p")
```

It's natural to use `str_count()` with `mutate()`:

```{r}
babynames %>% 
  count(name) %>% 
  mutate(
    vowels = str_count(name, "[aeiou]"),
    consonants = str_count(name, "[^aeiou]")
  )
```

```{r}
babynames %>% 
  count(name, wt = n) %>% 
  mutate(
    vowels = str_count(name, regex("[aeiouy]", ignore_case = TRUE)),
    consonants = str_count(name, regex("[^aeiouy]", ignore_case = TRUE)),
    ratio = vowels / consonants
  )
```

### Replace matches

`str_replace_all()` allow you to replace matches with new strings.
The simplest use is to replace a pattern with a fixed string:

```{r}
x <- c("apple", "pear", "banana")
str_replace_all(x, "[aeiou]", "-")
```

With `str_replace_all()` you can perform multiple replacements by supplying a named vector.
The name gives a regular expression to match, and the value gives the replacement.

```{r}
x <- c("1 house", "1 person has 2 cars", "3 people")
str_replace_all(x, c("1" = "one", "2" = "two", "3" = "three"))
```

`str_remove_all()` is a short cut for `str_replace_all(x, pattern, "")` --- it removes matching patterns from a string.

Use in `mutate()`

Using pipe inside mutate.
Recommendation to make a function, and think about testing it --- don't need formal tests, but useful to build up a set of positive and negative test cases as you.

### Exercises

1.  What word has the highest number of vowels?
    What word has the highest proportion of vowels?
    (Hint: what is the denominator?)

2.  For each of the following challenges, try solving it by using both a single regular expression, and a combination of multiple `str_detect()` calls.

    a.  Find all words that start or end with `x`.
    b.  Find all words that start with a vowel and end with a consonant.
    c.  Are there any words that contain at least one of each different vowel?

3.  Replace all forward slashes in a string with backslashes.

4.  Implement a simple version of `str_to_lower()` using `str_replace_all()`.

5.  Switch the first and last letters in `words`.
    Which of those strings are still `words`?

## Extract data from strings

Common for multiple variables worth of data to be stored in a single string.
In this section you'll learn how to use various functions tidyr to extract them.

Waiting on: <https://github.com/tidyverse/tidyups/pull/15>

## Locale dependent operations {#other-languages}

So far all of our examples have been using English.
The details of the many ways other languages are different to English are too diverse to detail here, but I wanted to give a quick outline of the functions who's behaviour differs based on your **locale**, the set of settings that vary from country to country.

-   Words are broken up by spaces.
-   Words are composed of individual spaces.
-   All letters in a word are written down.

The locale is usually a ISO 639 language code, which is a two or three letter abbreviation like "en" for English, "fr" for French, and "es" for Spanish.
If you don't already know the code for your language, [Wikipedia](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) has a good list, and you can see which are supported with `stringi::stri_locale_list()`.

Base R string functions automatically use your locale current locale, but stringr functions all default to the English locale.
This ensures that your code works the same way on every system, avoiding subtle bugs.
To choose a different locale you'll need to specify the `locale` argument; seeing that a function has a locale argument tells you that its behaviour will differ from locale to locale.

Here are a few places where locale matter:

-   Upper and lower case: only relatively few languages have upper and lower case (Latin, Greek, and Cyrillic, plus a handful of lessor known languages).
    The rules are not te same in every language that uses these alphabets.
    For example, Turkish has two i's: with and without a dot, and it has a different rule for capitalising them:

    ```{r}
    str_to_upper(c("i", "ı"))
    str_to_upper(c("i", "ı"), locale = "tr")
    ```

-   This also affects case insensitive matching with `coll(ignore_case =  TRUE)` which you can control with `coll()`:

    ```{r}
    i <- c("Iİiı")

    str_view_all(i, coll("i", ignore_case = TRUE))
    str_view_all(i, coll("i", ignore_case = TRUE, locale = "tr"))
    ```

-   Many characters with diacritics can be recorded in multiple ways: these will print identically but won't match with `fixed()`.

    ```{r}
    a1 <- "\u00e1"
    a2 <- "a\u0301"
    c(a1, a2)
    a1 == a2

    str_view(a1, fixed(a2))
    str_view(a1, coll(a2))
    ```

-   Another important operation that's affected by the locale is sorting.
    The base R `order()` and `sort()` functions sort strings using the current locale.
    If you want robust behaviour across different computers, you may want to use `str_sort()` and `str_order()` which take an additional `locale` argument.
    Here's an example: in Czech, "ch" is a digraph that appears after `h` in the alphabet.

    ```{r}
    str_sort(c("a", "c", "ch", "h", "z"))
    str_sort(c("a", "c", "ch", "h", "z"), locale = "cs")
    ```

    Danish has a similar problem.
    Normally, characters with diacritic sorts after the plain character.
    But in Danish ø and å are letters that come at the end of the alphabet:

    ```{r}
    str_sort(c("a", "å", "o", "ø", "z"))
    str_sort(c("a", "å", "o", "ø", "z"), locale = "da")
    ```

    TODO after dplyr 1.1.0: discuss `arrange()`