stringr str_split() in R: Split String by Pattern

The str_split() function in stringr divides each input string into pieces by a pattern, returning a LIST of character vectors. Use str_split_fixed() when you need a fixed-column matrix.

str_split(x, "/")                           # list (one element per string)
str_split(x, "/", n = 2)                    # split into at most 2 pieces
str_split_fixed(x, "/", n = 3)              # matrix with 3 columns
str_split_1(x, "/")                         # single string -> vector (stringr 1.5+)
str_split_i(x, "/", i = 2)                  # extract i-th piece
unlist(str_split(x, "/"))                   # flatten to single vector
str_split(x, regex("[,;]"))                 # multiple delimiters via regex

What str_split() does in one sentence

str_split(string, pattern) returns a LIST where each element is a character vector of pieces from the corresponding input string. This is necessary because different strings may produce different numbers of pieces.

For most uses, you immediately want one of: a vector (single input), a matrix (consistent piece count), or a specific piece (i-th element). stringr provides specialized functions for each case.

Syntax

str_split(string, pattern, n = Inf). Returns a list. Pattern is regex by default.

The result is a list of length 3 (one per input string). Each element is a character vector of pieces.

Five common patterns

1. Basic split

2. Split into matrix

The second row has only 2 pieces; the missing third becomes an empty string.

3. Single string -> vector

str_split_1() (stringr 1.5+) is for the common case of splitting ONE string. Returns a flat vector instead of a 1-element list.

4. Extract i-th piece

str_split_i() does split + indexing in one call. Cleaner than sapply(str_split(x, "-"), \(p) p[2]).

5. Multi-character or regex delimiter

Pass a regex to split on multiple delimiters. The character class [,;] matches either comma or semicolon.

Common pitfalls

Pitfall 1: forgetting that str_split() returns a LIST. Even for a single input, str_split("a/b", "/") returns a list of one element, not a vector. Use str_split_1() for vector output.

Pitfall 2: regex pattern when literal was intended. str_split(x, ".") splits on EVERY character (regex .). Use fixed(".") to split on actual dot.

Try it yourself

Related stringr functions

After mastering str_split, look at:

• tidyr::separate_wider_delim(): split a column in a data frame
• str_subset(): filter strings matching a pattern
• str_extract(): extract matched substrings
• str_count(): count delimiter occurrences (predict pieces)
• strsplit(): base R equivalent

For data frame columns, prefer tidyr::separate_wider_delim() over manually splitting and reshaping.

FAQ

How do I split a string by delimiter in R?

Use stringr::str_split(x, "delimiter") for general use (returns list). For a single string, str_split_1(x, "delimiter") returns a vector. For a matrix, str_split_fixed(x, "delimiter", n).

What is the difference between str_split and strsplit?

Both return a list of character vectors. str_split() is from stringr, more consistent (NA -> NA), and integrates better with the tidyverse. strsplit() is base R with no package dependency.

How do I split into a fixed number of pieces?

Use str_split_fixed(x, pattern, n = N) for a matrix with exactly N columns. Or str_split(x, pattern, n = N) to limit pieces in the list output (everything past N stays joined).

How do I get just one piece from each split?

Use str_split_i(x, pattern, i = 2) to extract the i-th piece from each split. Cleaner than mapping over the list output.

How do I split a column in a data frame?

Use tidyr::separate_wider_delim(df, cols = col, delim = "/", names = c("a","b","c")). It is the data-frame-specific tool, while str_split is for raw strings.