Implements drop_na()

NEWS.md

@@ -1,5 +1,9 @@
 # dtplyr (development version)
 
+* More tidyr verbs are in the process of being added:
+
+ * `drop_na()` (@markfairbanks, #194)
+
 # dtplyr 1.1.0
 
 ## New features
@@ -29,7 +33,7 @@
  (apart from predicate functions which can't easily work on lazily evaluated
  data tables).
 
-* We have begun the process of add translations for tidyr verbs beginning
+* We have begun the process of adding translations for tidyr verbs beginning
  with `pivot_wider()` (@markfairbanks, #189).
 
 ## Translation improvements

R/step-call.R

@@ -231,3 +231,41 @@ unique.dtplyr_step <- function(x, incomparables = FALSE, ...) {
  }
  distinct(x)
 }
+
+# tidyr verbs -------------------------------------------------------------
+#' Drop rows containing missing values
+#'
+#' @description
+#' This is a method for the tidyr `drop_na()` generic. It is translated to
+#' `data.table::na.omit()`
+#'
+#' @param data A [lazy_dt()].
+#' @inheritParams tidyr::drop_na
+#' @examples
+#' library(dplyr)
+#' library(tidyr)
+#'
+#' dt <- lazy_dt(tibble(x = c(1, 2, NA), y = c("a", NA, "b")))
+#' dt %>% drop_na()
+#' dt %>% drop_na(x)
+#'
+#' vars <- "y"
+#' dt %>% drop_na(x, any_of(vars))
+# exported onLoad
+drop_na.dtplyr_step <- function(data, ...) {
+ sim_data <- simulate_vars(data)
+ locs <- names(tidyselect::eval_select(expr(c(...)), sim_data))
+
+ args <- list()
+ if (length(locs) > 0) {
+ args$cols <- locs
+ }
+
+ step_call(data, "na.omit", args = args)
+}
+
+# exported onLoad
+drop_na.data.table <- function(data, ...) {
+ data <- lazy_dt(data)
+ tidyr::drop_na(data, ...)
+}

R/zzz.R

@@ -4,12 +4,14 @@
  register_s3_method("dplyr", "intersect", "data.table")
  register_s3_method("dplyr", "setdiff", "data.table")
  register_s3_method("dplyr", "union", "data.table")
+ register_s3_method("tidyr", "drop_na", "data.table")
  register_s3_method("tidyr", "pivot_wider", "data.table")
 
  register_s3_method("dplyr", "filter", "dtplyr_step")
  register_s3_method("dplyr", "intersect", "dtplyr_step")
  register_s3_method("dplyr", "setdiff", "dtplyr_step")
  register_s3_method("dplyr", "union", "dtplyr_step")
+ register_s3_method("tidyr", "drop_na", "dtplyr_step")
  register_s3_method("tidyr", "pivot_wider", "dtplyr_step")
 }
 

_pkgdown.yml

@@ -39,6 +39,7 @@ reference:
 
 - title: tidyr verbs
  contents:
+ - drop_na.dtplyr_step
  - pivot_wider.dtplyr_step
 
 navbar:

tests/testthat/_snaps/step-call.md

@@ -16,3 +16,11 @@
  Output
  setnames(copy(DT), c("a", "b", "c"), toupper)
 
+# errors are raised
+
+ Code
+ collect(drop_na(dt, "z"))
+ Error <vctrs_error_subscript_oob>
+ Can't subset columns that don't exist.
+ x Column `z` doesn't exist.
+

tests/testthat/test-step-call.R

@@ -161,3 +161,36 @@ test_that("unique is an alias for distinct", {
  dt <- lazy_dt(data.table(x = c(1, 1)))
  expect_equal(unique(dt), distinct(dt))
 })
+
+# drop_na ------------------------------------------------------------------
+
+test_that("empty call drops every row", {
+ tb <- tibble(x = c(1, 2, NA), y = c("a", NA, "b"))
+ step <- drop_na(lazy_dt(tb, "DT"))
+ expect_equal(show_query(step), expr(na.omit(DT)))
+ expect_equal(as_tibble(step), tb[1, ])
+})
+
+test_that("uses specified variables", {
+ df <- tibble(x = c(1, 2, NA), y = c("a", NA, "b"))
+ dt <- lazy_dt(df, "DT")
+
+ step <- drop_na(dt, x)
+ expect_equal(show_query(step), expr(na.omit(DT, cols = "x")))
+ expect_equal(collect(step), df[1:2, ])
+
+ step <- drop_na(dt, x:y)
+ expect_equal(show_query(step), expr(na.omit(DT, cols = !!c("x", "y"))))
+ expect_equal(collect(step), df[1, ])
+})
+
+test_that("errors are raised", {
+ tb <- tibble(x = c(1, 2, NA), y = c("a", NA, "b"))
+ dt <- lazy_dt(tb, "DT")
+ expect_snapshot(collect(drop_na(dt, "z")), error = TRUE)
+})
+
+test_that("converts data.table to dtplyr_step", {
+ df <- data.table(x = c(1, 2, NA), y = c("a", NA, "b"))
+ expect_s3_class(drop_na(df), "dtplyr_step_call")
+})