ARROW-16480: [R] Update read_csv_arrow and open_dataset parse_options, read_options, and convert_options to take lists

r/R/csv.R

@@ -207,6 +207,7 @@ read_delim_arrow <- function(file,
     }
     on.exit(file$close())
   }
+
   reader <- CsvTableReader$create(
     file,
     read_options = read_options,
@@ -326,6 +327,18 @@ CsvTableReader$create <- function(file,
                                   ...) {
   assert_is(file, "InputStream")
 
+  if (is.list(read_options)) {
+    read_options <- do.call(CsvReadOptions$create, read_options)
+  }
+
+  if (is.list(parse_options)) {
+    parse_options <- do.call(CsvParseOptions$create, parse_options)
+  }
+
+  if (is.list(convert_options)) {
+    convert_options <- do.call(CsvConvertOptions$create, convert_options)
+  }
+
   if (!(tolower(read_options$encoding) %in% c("utf-8", "utf8"))) {
     file <- MakeReencodeInputStream(file, read_options$encoding)
   }

r/R/dataset-format.R

@@ -136,6 +136,18 @@ CsvFileFormat$create <- function(...,
     ))
   }
 
+  if (!inherits(read_options, "CsvReadOptions")) {
+    read_options <- do.call(CsvReadOptions$create, read_options)
+  }
+
+  if (!inherits(convert_options, "CsvConvertOptions")) {
+    convert_options <- do.call(CsvConvertOptions$create, convert_options)
+  }
+
+  if (!inherits(opts, "CsvParseOptions")) {
+    opts <- do.call(CsvParseOptions$create, opts)
+  }
+
   column_names <- read_options$column_names
   schema_names <- names(schema)
 

r/tests/testthat/test-csv.R

@@ -664,3 +664,26 @@ test_that("Shows an error message when trying to read a timestamp with time zone
     "CSV conversion error to timestamp\\[ns\\]: expected no zone offset in"
   )
 })
+
+test_that("CSV reading/parsing/convert options can be passed in as lists", {
+  tf <- tempfile()
+  on.exit(unlink(tf))
+
+  writeLines('"x"\nNA\nNA\n"NULL"\n\n"foo"\n', tf)
+
+  tab1 <- read_csv_arrow(
+    tf,
+    convert_options = list(null_values = c("NA", "NULL"), strings_can_be_null = TRUE),
+    parse_options = list(ignore_empty_lines = FALSE),
+    read_options = list(skip_rows = 1L)
+  )
+
+  tab2 <- read_csv_arrow(
+    tf,
+    convert_options = CsvConvertOptions$create(null_values = c(NA, "NA", "NULL"), strings_can_be_null = TRUE),
+    parse_options = CsvParseOptions$create(ignore_empty_lines = FALSE),
+    read_options = CsvReadOptions$create(skip_rows = 1L)
+  )
+
+  expect_equal(tab1, tab2)
+})

r/tests/testthat/test-dataset-csv.R

@@ -444,11 +444,35 @@ test_that("skip argument in open_dataset", {
 })
 
 test_that("error message if non-schema passed in as schema to open_dataset", {
-
   # passing in the schema function, not an actual schema
   expect_error(
     open_dataset(csv_dir, format = "csv", schema = schema),
     regexp = "`schema` must be an object of class 'Schema' not 'function'.",
     fixed = TRUE
   )
 })
+
+test_that("CSV reading/parsing/convert options can be passed in as lists", {
+  tf <- tempfile()
+  on.exit(unlink(tf))
+
+  writeLines('"x"\n"y"\nNA\nNA\n"NULL"\n\n"foo"\n', tf, )
+
+  ds1 <- open_dataset(
+    tf,
+    format = "csv",
+    convert_options = list(null_values = c("NA", "NULL"), strings_can_be_null = TRUE),
+    read_options = list(skip_rows = 1L)
+  ) %>%
+    collect()
+
+  ds2 <- open_dataset(
+    tf,
+    format = "csv",
+    convert_options = CsvConvertOptions$create(null_values = c(NA, "NA", "NULL"), strings_can_be_null = TRUE),
+    read_options = CsvReadOptions$create(skip_rows = 1L)
+  ) %>%
+    collect()
+
+  expect_equal(ds1, ds2)
+})