Add new function 'to_df' which is the same as 'to_list' but return data.frame (issue #2)

Author: gdemin

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: comprehenr
 Type: Package
 Title: List Comprehensions
-Version: 0.6.9
+Version: 0.6.10
 Maintainer: Gregory Demin <gdemin@gmail.com>
 Authors@R: person("Gregory", "Demin", email = "gdemin@gmail.com",
          role = c("aut", "cre"))

NAMESPACE

@@ -6,6 +6,7 @@ export(exclude)
 export(lag_list)
 export(mark)
 export(numerate)
+export(to_df)
 export(to_list)
 export(to_vec)
 export(unmark)

NEWS

@@ -1,3 +1,7 @@
+0.6.10 (31-01-2021)
+================
+* add new function 'to_df' which combine result of list comprehension to data.frame (issue #2)
+
 0.6.9 (18-01-2021)
 ================
 * fix bug with nested 'to_list'/'to_vec'/'alter' (issue #1)

R/to_list.R

@@ -21,6 +21,8 @@ SPECIAL_FUNCTIONS = expression(
 #' `repeat`. You can iterate over multiple lists if you provide several
 #' loop variables in backticks. See examples.
 #' - `to_vec` is the same as 'to_list' but return vector. See examples.
+#' - `to_df` is the same as 'to_list' but return data.frame. All elements of
+#' resulted list will be converted to data.frame and combined via `rbind`.
 #' - `alter` returns the same type as its argument but with modified
 #' elements. It is useful for altering existing data.frames or lists. See
 #' examples.
@@ -30,6 +32,7 @@ SPECIAL_FUNCTIONS = expression(
 #' @param expr expression which starts with `for`, `while` or `repeat`.
 #' @param recursive	logical. Should unlisting be applied to list components of result? See [unlist][base::unlist] for details.
 #' @param use.names logical. Should names be preserved? See [unlist][base::unlist] for details.
+#' @param fill logical. TRUE by default. Should we combine data.frames with different names in the `to_df`?
 #' @param data data.frame/list/vector which we want to alter
 #' @return list for `to_list` and vector for `to_vec`
 #' @export
@@ -58,6 +61,9 @@ SPECIAL_FUNCTIONS = expression(
 #' # gives only locally increasing values
 #' to_vec(for(`i, j` in lag_list(rand_sequence)) if(j>i) j)
 #'
+#' # to_df
+#' to_df(for(`name, x` in mark(mtcars)) list(mean = mean(x), sd = sd(x), var = name))
+#'
 #' # 'alter' examples
 #' data(iris)
 #' # scale numeric variables
@@ -248,6 +254,49 @@ alter = function(expr, data = NULL){
 }
 
 
+#' @rdname to_list
+#' @export
+to_df = function(expr, fill = TRUE){
+    res = eval.parent(
+        substitute(
+            comprehenr::to_list(expr)
+        )
+    )
+    for(i in seq_along(res)){
+        if(!is.null(res[[i]]) && !is.data.frame(res[[i]])){
+            res[[i]] = as.data.frame(as.list(res[[i]]))
+        }
+    }
+    #
+    if(fill){
+        all_names = unique(unlist(lapply(res, names), use.names = FALSE))
+        for(i in seq_along(res)){
+            new_names = setdiff(all_names, names(res[[i]]))
+            if(length(new_names)>0){
+                res[[i]][,new_names] = NA
+            }
+        }
+    }
+
+    do.call(rbind, res)
+}
+
+# @rdname to_list
+# @export
+# to_dfr = to_df
+
+# @rdname to_list
+# @export
+# to_dfc = function(expr = NULL){
+#     res = eval.parent(
+#         substitute(
+#             comprehenr::to_list(expr)
+#         )
+#     )
+#     as.data.frame(res)
+# }
+
+
 #' @rdname to_list
 #' @export
 exclude = function(){

inst/tinytest/test_to_list.R

@@ -217,3 +217,32 @@ mtcars2= alter(for(`i,j` in lag_list(mtcars)) if(length(unique(i))==2) exclude()
 res_mtcars = mtcars
 res_mtcars[, c("am", "vs")]  = NULL
 expect_equal(mtcars2, res_mtcars)
+
+context("to_df/to_dfc")
+
+data("mtcars")
+
+expect_identical(
+    to_df(for(each in list(1:2, 2:3, 3:4)) each),
+    structure(list(X1L = c(1L, NA, NA), X2L = c(2L, 2L, NA), X3L = c(NA,
+                                                                     3L, 3L), X4L = c(NA, NA, 4L)), row.names = c(NA, -3L), class = "data.frame")
+)
+
+expect_error(
+    to_df(for(each in list(1:2, 2:3, 3:4)) each, fill = FALSE),
+)
+
+expect_equal(
+    to_df(for(`.name, .x` in mark(mtcars)) list(mean = mean(.x), sd = sd(.x), var = .name)),
+    data.frame(mean = colMeans(mtcars), sd = sapply(mtcars, sd), var = names(mtcars), row.names = seq_along(mtcars))
+)
+
+
+expect_identical(
+    to_df(for(`.index, .x` in enumerate(mtcars)) list(mean = mean(.x), sd = sd(.x), item_id = .index)),
+    data.frame(mean = colMeans(mtcars), sd = sapply(mtcars, sd), item_id = seq_along(mtcars), row.names = seq_along(mtcars))
+)
+
+
+
+