Add negate argument to str_subset(), str_detect(), and str_which()

NEWS.md

@@ -3,6 +3,10 @@
 * `str_interp()` now renders lists consistently independent on the presence of
   additional placeholders (@amhrasmussen)
 
+* `str_subset()`, `str_detect()`, and `str_which()` gets `negate` argument,
+  which is useful when you want the elements that do NOT match (#259,
+  @yutannihilation).
+
 # stringr 1.3.1
 
 * `str_replace_all()` with a named vector now respects modifier functions (#207)

R/detect.r

@@ -20,6 +20,8 @@
 #'   Match character, word, line and sentence boundaries with
 #'   [boundary()]. An empty pattern, "", is equivalent to
 #'   `boundary("character")`.
+#'
+#' @param negate If `TRUE`, return non-matching elements.
 #' @return A logical vector.
 #' @seealso [stringi::stri_detect()] which this function wraps,
 #'   [str_subset()] for a convenient wrapper around
@@ -35,12 +37,15 @@
 #'
 #' # Also vectorised over pattern
 #' str_detect("aecfg", letters)
-str_detect <- function(string, pattern) {
+#'
+#' # Returns TRUE if the pattern do NOT match
+#' str_detect(fruit, "^p", negate = TRUE)
+str_detect <- function(string, pattern, negate = FALSE) {
   switch(type(pattern),
     empty = ,
-    bound = str_count(string, pattern) > 0,
-    fixed = stri_detect_fixed(string, pattern, opts_fixed = opts(pattern)),
-    coll  = stri_detect_coll(string,  pattern, opts_collator = opts(pattern)),
-    regex = stri_detect_regex(string, pattern, opts_regex = opts(pattern))
+    bound = str_count(string, pattern) > 0 & !negate,
+    fixed = stri_detect_fixed(string, pattern, negate = negate, opts_fixed = opts(pattern)),
+    coll  = stri_detect_coll(string,  pattern, negate = negate, opts_collator = opts(pattern)),
+    regex = stri_detect_regex(string, pattern, negate = negate, opts_regex = opts(pattern))
   )
 }

R/subset.R

@@ -23,21 +23,24 @@
 #' str_subset(fruit, "b")
 #' str_subset(fruit, "[aeiou]")
 #'
+#' # Returns elements that do NOT match
+#' str_subset(fruit, "^p", negate = TRUE)
+#'
 #' # Missings never match
 #' str_subset(c("a", NA, "b"), ".")
 #' str_which(c("a", NA, "b"), ".")
-str_subset <- function(string, pattern) {
+str_subset <- function(string, pattern, negate = FALSE) {
   switch(type(pattern),
     empty = ,
-    bound = string[str_detect(string, pattern)],
-    fixed = stri_subset_fixed(string, pattern, omit_na = TRUE, opts_fixed = opts(pattern)),
-    coll  = stri_subset_coll(string, pattern, omit_na = TRUE, opts_collator = opts(pattern)),
-    regex = stri_subset_regex(string, pattern, omit_na = TRUE, opts_regex = opts(pattern))
+    bound = string[str_detect(string, pattern) & !negate],
+    fixed = stri_subset_fixed(string, pattern, omit_na = TRUE, negate = negate, opts_fixed = opts(pattern)),
+    coll  = stri_subset_coll(string, pattern, omit_na = TRUE, negate = negate, opts_collator = opts(pattern)),
+    regex = stri_subset_regex(string, pattern, omit_na = TRUE, negate = negate, opts_regex = opts(pattern))
   )
 }
 
 #' @export
 #' @rdname str_subset
-str_which <- function(string, pattern) {
-  which(str_detect(string, pattern))
+str_which <- function(string, pattern, negate = FALSE) {
+  which(str_detect(string, pattern, negate = negate))
 }

tests/testthat/test-detect.r

@@ -8,6 +8,9 @@ test_that("special cases are correct", {
 test_that("vectorised patterns work", {
   expect_equal(str_detect("ab", c("a", "b", "c")), c(T, T, F))
   expect_equal(str_detect(c("ca", "ab"), c("a", "c")), c(T, F))
+
+  # negation works
+  expect_equal(str_detect("ab", c("a", "b", "c"), negate = TRUE), c(F, F, T))
 })
 
 test_that("modifiers work", {

tests/testthat/test-subset.r

@@ -6,11 +6,20 @@ test_that("basic subsetting for fixed patterns works", {
     str_subset(c("i", "I"), fixed("i", ignore_case = TRUE)),
     c("i", "I")
   )
+
+  # negation works
+  expect_equal(str_subset(c("i", "I"), fixed("i"), negate = TRUE), "I")
 })
 
 test_that("str_which is equivalent to grep", {
   expect_equal(
     str_which(head(letters), "[aeiou]"),
     grep("[aeiou]", head(letters))
   )
+
+  # negation works
+  expect_equal(
+    str_which(head(letters), "[aeiou]", negate = TRUE),
+    grep("[aeiou]", head(letters), invert = TRUE)
+  )
 })