16 Strings: Creation, Concatenation, and Substrings

What This Chapter Covers

This chapter is a thorough introduction to working with text in R. You will learn how character values are created and stored in character vectors, the two styles of quoting (" and '), the common escape sequences (\n, \t, \", \\), and the length metrics length() and nchar(), which do not mean the same thing. You will see the four everyday concatenation tools (paste(), paste0(), cat(), sprintf()), the substring-extraction functions (substr(), substring(), startsWith(), endsWith()), and how to change case with toupper(), tolower(), tools::toTitleCase(), and trimws(). The chapter sticks to base R; Chapter 17 covers the richer stringr toolkit and regular expressions. By the end of this chapter you will be able to build, inspect, extract, and clean text cleanly.

flowchart LR
    T["Text in R"] --> CR["Create <br> 'hello' / character(n) / readLines()"]
    T --> CO["Concatenate <br> paste / paste0 / cat / sprintf"]
    T --> SU["Substring <br> substr / substring / startsWith / endsWith"]
    T --> CL["Clean <br> toupper / tolower / trimws"]
    style T fill:#e3f2fd,stroke:#1976D2
    style CR fill:#fff3e0,stroke:#F57C00
    style CO fill:#fff3e0,stroke:#F57C00
    style SU fill:#f3e5f5,stroke:#8E24AA
    style CL fill:#e8f5e9,stroke:#388E3C

16.1 Character Values and Character Vectors

Core Concept: Every String Is a One-Element Character Vector

R has no separate “character” and “string” types. A single letter and a full paragraph are both character vectors of length 1. Multiple strings sit side by side in a longer character vector.

Common Mistake: Confusing length() and nchar()

length(x) reports how many strings the vector holds. nchar(x) reports how many characters are inside each string. They answer different questions and are the source of a classic beginner bug.

16.2 Quotes and Escape Sequences

Double vs Single Quotes

Both styles delimit a string; pick whichever avoids escaping. The R community and tidyverse style guide prefer double quotes by default.

Common Escape Sequences

Sequence	Meaning
`\n`	Newline.
`\t`	Tab.
`\\`	A literal backslash.
`\"` / `\'`	A literal double / single quote inside the matching delimiter.
`\u00e9`	A Unicode code point (here, é).

Expert Insight: print() Shows Literals, cat() Renders Them

When print() displays a string, it shows the escape sequences as you typed them. When cat() or message() writes the string out, it interprets those escapes. This trips up beginners who think their \n isn’t working; the answer is usually that they are looking at print() output.

16.3 Building Strings Programmatically

paste() and paste0()

paste() joins pieces with a separator (default one space); paste0() is the same with sep = "". Both are vectorised and recycle a length-1 argument against a longer one.

paste() With collapse

paste() has a second argument, collapse, that joins a vector into a single string after concatenation. This is the right way to turn a character vector into one line of text.

sprintf() for Precise Formatting

sprintf() uses C-style format specifiers for controlled width, padding, and decimals. You saw it in Chapter 5; it is the right tool whenever the exact visual layout matters.

Best Practice: Pick the Right Tool

Reach for paste() when joining strings with a delimiter, paste0() when building file names or URLs, and sprintf() whenever decimals, widths, or padding matter. Reserve cat() for writing a message to the console rather than building a string.

16.4 Length Metrics

Everyday Measurements

Unicode, Bytes, and nchar()

nchar() reports characters by default, not bytes. A string with accented letters or emojis takes more bytes than characters. You can ask for a byte count with nchar(x, type = "bytes") when it matters for file size estimates.

16.5 Extracting Substrings

substr(): Positions as Start and Stop

substr(x, start, stop) returns the slice from position start to position stop, inclusive on both ends, 1-indexed. It is vectorised.

substring(): Allows Multiple Start Points

substring() is the older cousin. It defaults first = 1 and last = 1000000L and, usefully, supports multiple start positions at once, producing several substrings from a single input string.

Replacing Substrings with substr<-

Assigning to substr(x, start, stop) replaces the slice in place. The replacement must be the same length as the slice; if shorter, R truncates; if longer, only the first characters are used.

16.6 Prefix, Suffix, and Case

startsWith() and endsWith() for Prefix and Suffix Tests

Both functions are vectorised and return a logical vector.

Changing Case

Four functions cover the everyday cases.

Function	Purpose
`toupper(x)`	Convert to uppercase.
`tolower(x)`	Convert to lowercase.
`tools::toTitleCase(x)`	Title Case.
`chartr(old, new, x)`	Fixed-position character translation (1-to-1).

Expert Insight: Normalise Case at the Boundary

Before any string comparison or grouping, decide on a canonical case (often lowercase) and apply it once at the point where the data enters your program. “New Delhi”, “new delhi”, and “NEW DELHI” are three different values until you normalise them.

16.7 Removing Whitespace

trimws() Strips Leading and Trailing Spaces

trimws() has three modes: both ends ("both", the default), only the left, or only the right.

Common Mistake: Forgetting That trimws() Only Touches the Ends

Double spaces in the middle of a string survive trimws(). If you need to collapse them, use gsub("\\s+", " ", x) (covered in Chapter 17).

16.8 Splitting Strings with `strsplit()`

From One String to Several

strsplit(x, split) breaks each string on the separator and returns a list with one element per input string. This reflects that different strings may split into different numbers of pieces.

Best Practice: strsplit() Is for Unknown-Length Cases

If you know every string splits into the same number of fields (for example, well-formed fixed-column text), do.call(rbind, strsplit(...)) gives you a matrix. For real CSV data, use read.csv() or readr::read_csv() rather than hand-splitting strings.

16.9 A Worked Example: Tidying a Messy Name List

Putting the Chapter Together

Every core technique of the chapter appears: trimws() to drop edge whitespace, case normalisation with tolower() and toTitleCase(), strsplit() to break each name into words, substr() to pull first characters, and paste0() with collapse to assemble the initials.

16.10 Summary

Summary of concepts introduced in this chapter
Concept	Description
Foundations
Strings Are Character Vectors	A single string is a character vector of length 1
Quotes and Escapes	Prefer double quotes; \n, \t, \\, \" are the common escapes
length() vs nchar()	length() counts strings; nchar() counts characters inside each string
Build and Format
paste() and paste0()	Join pieces into strings; the collapse argument folds a vector into one string
sprintf()	Precise formatting for widths, decimals, and padding using C-style specifiers
Slice and Test
substr() and substring()	Slice by start and stop positions; both are vectorised
startsWith() and endsWith()	Vectorised prefix and suffix tests returning logical vectors
Clean and Split
Case Conversion	toupper, tolower, tools::toTitleCase normalise case
trimws()	Strips leading and trailing whitespace from each element
strsplit()	Splits each string on a separator; returns a list, one vector per input

Applying This in Practice

Text is the messiest kind of data you will handle, and the cost of skipping a cleanup step compounds rapidly downstream. Make trimws(), case normalisation, and consistent separators a routine first pass on any character column. The next chapter extends these tools to the richer world of regular expressions, pattern matching, and the stringr package, the toolkit you need as soon as simple position-based slicing stops being enough.