Adding keep_empty argument to list_c, list_cbind, and list_rbind.

NEWS.md

@@ -1,5 +1,7 @@
 # purrr (development version)
 
+* Added a new keep_empty argument to `list_c()`, `list_cbind()`, and `list_rbind()` which will keep empty elements as `NA` in the returned data frame (@SokolovAnatoliy, #1096).
+
 # purrr 1.0.2
 
 * Fixed valgrind issue.

R/list-combine.R

@@ -22,6 +22,9 @@
 #'   same size (i.e. number of rows).
 #' @param name_repair One of `"unique"`, `"universal"`, or `"check_unique"`.
 #'   See [vctrs::vec_as_names()] for the meaning of these options.
+#' @param keep_empty  An optional logical. If `FALSE` (the default), then
+#'   empty (`NULL`) elements are silently ignored; if `TRUE`, then empty
+#'   elements are preserved by converting to `NA`.
 #' @inheritParams rlang::args_dots_empty
 #' @export
 #' @examples
@@ -30,16 +33,21 @@
 #'
 #' x2 <- list(
 #'   a = data.frame(x = 1:2),
-#'   b = data.frame(y = "a")
+#'   b = data.frame(y = "a"),
+#'   c = NULL
 #' )
 #' list_rbind(x2)
 #' list_rbind(x2, names_to = "id")
+#' list_rbind(x2, names_to = "id", keep_empty = TRUE)
 #' list_rbind(unname(x2), names_to = "id")
-#'
 #' list_cbind(x2)
-list_c <- function(x, ..., ptype = NULL) {
+#' list_cbind(x2, keep_empty = TRUE)
+list_c <- function(x, ..., ptype = NULL, keep_empty = FALSE) {
   vec_check_list(x)
   check_dots_empty()
+  if (keep_empty) {
+    x <- convert_null_to_NA(x)
+  }
 
   # For `list_c()`, we don't expose `list_unchop()`'s `name_spec` arg,
   # and instead strip outer names to avoid collisions with inner names
@@ -58,19 +66,26 @@ list_cbind <- function(
     x,
     ...,
     name_repair = c("unique", "universal", "check_unique"),
-    size = NULL
+    size = NULL,
+    keep_empty = FALSE
   ) {
   check_list_of_data_frames(x)
   check_dots_empty()
+  if (keep_empty) {
+    x <- convert_null_to_NA(x)
+  }
 
   vec_cbind(!!!x, .name_repair = name_repair, .size = size, .error_call = current_env())
 }
 
 #' @export
 #' @rdname list_c
-list_rbind <- function(x, ..., names_to = rlang::zap(), ptype = NULL) {
+list_rbind <- function(x, ..., names_to = rlang::zap(), ptype = NULL, keep_empty = FALSE) {
   check_list_of_data_frames(x)
   check_dots_empty()
+  if (keep_empty) {
+    x <- convert_null_to_NA(x)
+  }
 
   vec_rbind(!!!x, .names_to = names_to, .ptype = ptype, .error_call = current_env())
 }
@@ -95,3 +110,9 @@ check_list_of_data_frames <- function(x, error_call = caller_env()) {
     call = error_call
   )
 }
+
+convert_null_to_NA <- function(x) {
+  is_null <- map_lgl(x, is.null)
+  x[is_null] <- list(NA)
+  x
+}

tests/testthat/test-list-combine.R

@@ -70,6 +70,24 @@ test_that("NULLs are ignored", {
   expect_equal(list_cbind(list(df1, NULL, df2)), vec_cbind(df1, df2))
 })
 
+test_that("NULLs are converted to NA when keep_empty = TRUE", {
+  df1 <- data.frame(x = 1)
+  df2 <- data.frame(y = 1)
+
+  expect_equal(
+    list_c(list(1, NULL, 2), keep_empty = TRUE),
+    c(1, NA, 2)
+  )
+  expect_equal(
+    list_rbind(list(df1, NULL, df1), keep_empty = TRUE),
+    data.frame(x = c(1, NA, 1))
+  )
+  expect_equal(
+    list_cbind(list(df1, z = NULL, df2), keep_empty = TRUE),
+    data.frame(df1, z = NA, df2)
+  )
+})
+
 test_that("empty inputs return expected output", {
   expect_equal(list_c(list()), NULL)
   expect_equal(list_c(list(NULL)), NULL)