ministr

The purpose of ministr is to provide a zero-dependency reimplementation of stringr’s basic string manipulation verbs from the stringr package: padding, trimming, whitespace squishing, substrings, length, case conversion, duplication, concatenation, and wrapping. The README provides scope notes; this page walks through each function on a small set of toy strings, including two behaviours worth knowing about before you rely on it: how NA is propagated through concatenation, and how title case is found without a Unicode-aware word-break algorithm. (For pattern-matching functions like str_detect()/ str_replace(), see the companion minirx vignette.)

Functions

The ministr.R script supplies twelve core functions (grouped into eleven bullets below, since .str_to_upper()/.str_to_lower() share one):

• .str_pad(string, width, side, pad) is used to pad a string to a fixed width.
• .str_trim(string, side) is used to trim leading/trailing whitespace.
• .str_squish(string) is used to trim and collapse internal whitespace runs to a single space.
• .str_sub(string, start, end) is used to extract a substring by (possibly negative) character position.
• .str_length(string) is used to count the characters in a string.
• .str_to_upper(string) and .str_to_lower(string) convert case.
• .str_to_title(string) converts to title case.
• .str_to_sentence(string) converts to sentence case.
• .str_dup(string, times) is used to repeat a string.
• .str_c(..., sep, collapse) is used to concatenate strings, propagating NA.
• .str_wrap(string, width, indent, exdent) is used to wrap a string to a target line width.

It contains no internal helpers.

Padding

The .str_pad() function pads to a fixed width, left by default:

.str_pad(c("7", "70", "700"), width = 4, pad = "0")

[1] "0007" "0070" "0700"

Padding can go on the right, or on both sides at once (with any leftover padding placed on the right when the total isn’t evenly split):

.str_pad("hi", width = 6, side = "right", pad = "-")

[1] "hi----"

.str_pad("hi", width = 7, side = "both", pad = "-")

[1] "--hi---"

A string already at least as wide as width is returned unchanged, and an NA input stays NA rather than being padded. The rest of this page reuses this small toy vector, mixing untrimmed and inconsistently cased strings with a missing value:

names <- c("  bilbo baggins", "FRODO", NA)
.str_pad(names, width = 20)

[1] "       bilbo baggins" "               FRODO" NA                    

Trimming and squishing whitespace

The .str_trim() function trims both sides by default, or a single side on request:

.str_trim(names[1])

[1] "bilbo baggins"

.str_trim(names[1], side = "left")

[1] "bilbo baggins"

The .str_squish() function goes further, also collapsing any run of internal whitespace down to a single space:

.str_squish("  too   much   space  ")

[1] "too much space"

Substrings

The .str_sub() function extracts by character position, and accepts negative positions counting back from the end of the string, the same way stringr::str_sub() does:

.str_sub("hello world", 1, 5)

[1] "hello"

.str_sub("hello world", -5, -1)

[1] "world"

Positions past either end of the string are clamped rather than erroring, and a start after the end just gives an empty string:

.str_sub("hi", 1, 10)

[1] "hi"

.str_sub("hello", 4, 2)

[1] ""

Length

The .str_length() function counts characters, and an NA input gives an NA count rather than treating it as the two-character string "NA":

.str_length(names)

[1] 15  5 NA

Case conversion

The .str_to_upper()/.str_to_lower() functions are plain upper/lower casing:

.str_to_upper("Bilbo")

[1] "BILBO"

.str_to_lower("Bilbo")

[1] "bilbo"

The .str_to_title() function lower-cases the whole string first, then upper-cases the first letter of each word:

.str_to_title("  bilbo baggins")

[1] "  Bilbo Baggins"

The .str_to_sentence() function is similar, but only upper-cases the very first letter of the string – the rest is lower-cased regardless of word boundaries, since the whole input is treated as a single sentence rather than split on ./!/?:

.str_to_sentence("BILBO BAGGINS the hobbit")

[1] "Bilbo baggins the hobbit"

Word starts for .str_to_title() are found with the ASCII regex \\b, not a Unicode-aware word-break algorithm. That’s a fine approximation for whitespace-delimited Latin text, but it isn’t a substitute for stringi’s locale-aware title casing if your strings involve other scripts.

Duplication

The .str_dup() function repeats each string, recycling times against string:

.str_dup("ab", 3)

[1] "ababab"

.str_dup(c("x", "y"), c(2, 4))

[1] "xx"   "yyyy"

Concatenation and NA propagation

The .str_c() function behaves like paste0() with a separator, but an NA anywhere in the inputs to a given position makes that position’s result NA:

.str_c("Mr. ", names, sep = "")

[1] "Mr.   bilbo baggins" "Mr. FRODO"           NA                   

Compare that with what base paste0() does with the same input – it coerces the missing name to the literal string "NA" instead of propagating the missing value:

paste0("Mr. ", names)

[1] "Mr.   bilbo baggins" "Mr. FRODO"           "Mr. NA"             

Supplying collapse joins everything into one string, and an NA anywhere in the collapsed inputs makes the whole collapsed result NA:

.str_c(c("bilbo", "frodo"), collapse = ", ")

[1] "bilbo, frodo"

.str_c(c("bilbo", NA), collapse = ", ")

[1] NA

Wrapping

The .str_wrap() function rewraps a string to a target line width, returning a single string per input with lines joined by "\n":

wrapped <- .str_wrap(
  "Bilbo Baggins was a very well-to-do hobbit who lived in a hole in the ground.",
  width = 20
)
cat(wrapped)

Bilbo Baggins was a
very well-to-do
hobbit who lived in
a hole in the
ground.

Scope: no locale awareness

None of these functions are locale-aware, unlike their stringr counterparts (which delegate to stringi/ICU under the hood). That’s a deliberate trade-off: it keeps this mini free of a regex-engine or ICU dependency, at the cost of correctness on non-Latin scripts and locale-specific casing rules (Turkish dotless i, for example). If your strings need that, this mini isn’t a substitute for stringr itself.