GH-35949: [R] CSV File reader options class objects should print the …

…selected values (#35955) Fixes #35949 * Closes: #35949 Authored-by: Nic Crane <thisisnic@gmail.com> Signed-off-by: Nic Crane <thisisnic@gmail.com>
apache · Jun 11, 2023 · fe60932 · fe60932
1 parent 7887605
commit fe60932
Show file tree

Hide file tree

Showing 5 changed files with 132 additions and 2 deletions.
diff --git a/r/R/arrowExports.R b/r/R/arrowExports.R
diff --git a/r/R/csv.R b/r/R/csv.R
@@ -464,10 +464,25 @@ CsvTableReader$create <- function(file,
 CsvReadOptions <- R6Class("CsvReadOptions",
   inherit = ArrowObject,
   public = list(
-    encoding = NULL
+    encoding = NULL,
+    print = function(...) {
+      cat("CsvReadOptions\n")
+      for (attr in c(
+        "column_names", "block_size", "skip_rows", "autogenerate_column_names",
+        "use_threads", "skip_rows_after_names", "encoding"
+      )) {
+        cat(sprintf("%s: %s\n", attr, self[[attr]]))
+      }
+      invisible(self)
+    }
   ),
   active = list(
-    column_names = function() csv___ReadOptions__column_names(self)
+    column_names = function() csv___ReadOptions__column_names(self),
+    block_size = function() csv___ReadOptions__block_size(self),
+    skip_rows = function() csv___ReadOptions__skip_rows(self),
+    autogenerate_column_names = function() csv___ReadOptions__autogenerate_column_names(self),
+    use_threads = function() csv___ReadOptions__use_threads(self),
+    skip_rows_after_names = function() csv___ReadOptions__skip_rows_after_names(self)
   )
 )
 CsvReadOptions$create <- function(use_threads = option_use_threads(),
@@ -491,6 +506,7 @@ CsvReadOptions$create <- function(use_threads = option_use_threads(),
   )
 
   options$encoding <- encoding
+
   options
 }
 

diff --git a/r/src/arrowExports.cpp b/r/src/arrowExports.cpp
diff --git a/r/src/csv.cpp b/r/src/csv.cpp
@@ -77,6 +77,36 @@ SEXP csv___ReadOptions__column_names(
   return cpp11::as_sexp(options->column_names);
 }
 
+// [[arrow::export]]
+SEXP csv___ReadOptions__block_size(
+    const std::shared_ptr<arrow::csv::ReadOptions>& options) {
+  return cpp11::as_sexp(options->block_size);
+}
+
+// [[arrow::export]]
+SEXP csv___ReadOptions__skip_rows(
+    const std::shared_ptr<arrow::csv::ReadOptions>& options) {
+  return cpp11::as_sexp(options->skip_rows);
+}
+
+// [[arrow::export]]
+SEXP csv___ReadOptions__autogenerate_column_names(
+    const std::shared_ptr<arrow::csv::ReadOptions>& options) {
+  return cpp11::as_sexp(options->autogenerate_column_names);
+}
+
+// [[arrow::export]]
+SEXP csv___ReadOptions__use_threads(
+    const std::shared_ptr<arrow::csv::ReadOptions>& options) {
+  return cpp11::as_sexp(options->use_threads);
+}
+
+// [[arrow::export]]
+SEXP csv___ReadOptions__skip_rows_after_names(
+    const std::shared_ptr<arrow::csv::ReadOptions>& options) {
+  return cpp11::as_sexp(options->skip_rows_after_names);
+}
+
 // [[arrow::export]]
 std::shared_ptr<arrow::csv::ConvertOptions> csv___ConvertOptions__initialize(
     cpp11::list options) {

diff --git a/r/tests/testthat/test-dataset-csv.R b/r/tests/testthat/test-dataset-csv.R
@@ -574,3 +574,22 @@ test_that("open_delim_dataset params passed through to open_dataset", {
 
   expect_equal(ds$time, "16-01-2023")
 })
+
+test_that("CSVReadOptions printing", {
+  default_read_options <- CsvReadOptions$create()
+  custom_read_options <- CsvReadOptions$create(skip_rows = 102)
+
+  expect_output(print(default_read_options), "skip_rows: 0")
+  expect_output(print(custom_read_options), "skip_rows: 102")
+})
+
+test_that("CSVReadOptions field access", {
+  options <- CsvReadOptions$create()
+  expect_equal(options$skip_rows, 0)
+  expect_equal(options$autogenerate_column_names, FALSE)
+  expect_equal(options$skip_rows_after_names, 0)
+  expect_equal(options$use_threads, option_use_threads())
+  expect_equal(options$column_names, character(0))
+  expect_equal(options$block_size, 1048576L)
+  expect_equal(options$encoding, "UTF-8")
+})