ARROW-13615 [R] Bindings for stringr::str_to_sentence

r/R/dplyr-functions.R

@@ -341,6 +341,7 @@ arrow_string_join_function <- function(null_handling, null_replacement = NULL) {
 #   str_to_lower
 #   str_to_upper
 #   str_to_title
+#   str_to_sentence
 #
 # Arrow locale will be supported with ARROW-14126
 stop_if_locale_provided <- function(locale) {
@@ -367,6 +368,11 @@ nse_funcs$str_to_title <- function(string, locale = "en") {
   Expression$create("utf8_title", string)
 }
 
+nse_funcs$str_to_sentence <- function(string, locale = "en") {
+  stop_if_locale_provided(locale)
+  Expression$create("utf8_capitalize", string)
+}
+
 nse_funcs$str_trim <- function(string, side = c("both", "left", "right")) {
   side <- match.arg(side)
   trim_fun <- switch(side,

r/R/expression.R

@@ -53,6 +53,7 @@
   # str_to_lower is defined in dplyr-functions.R
   # str_to_title is defined in dplyr-functions.R
   # str_to_upper is defined in dplyr-functions.R
+  # str_to_sentence is defined in dplyr-functions.R
   # str_trim is defined in dplyr-functions.R
   "stri_reverse" = "utf8_reverse",
   # substr is defined in dplyr-functions.R

r/tests/testthat/test-dplyr-funcs-string.R

@@ -1336,3 +1336,27 @@ test_that("str_starts, str_ends, startsWith, endsWith", {
     df
   )
 })
+
+test_that("str_to_sentence", {
+  df <- tibble(
+    one_sent = c("first word", "the second word", "the third word"),
+    two_sent = c("first word. second word? third word! fourth word",
+                 "second word", "third word")
+  )
+
+  expect_dplyr_equal(
+    input %>%
+      mutate(sentence_case = str_to_sentence(one_sent)) %>%
+      collect(),
+    df
+  )
+
+  # there is something strange going on with str_to_sentence in stringr where
+  # it doesn't recognise `.` as a sentence end
+  expect_dplyr_error(
+    input %>%
+      mutate(sentence_case_two = str_to_sentence(two_sent)) %>%
+      collect(),
+    df
+  )
+})