BUG FIX: CSV Writer ignores the header parameter when no metadata is …

…provided (#8740)

Fixes: #6669

Authors:
  - Sheilah Kirui (https://github.com/skirui-source)

Approvers:
  - Vukasin Milovanovic (https://github.com/vuule)
  - David Wendt (https://github.com/davidwendt)

URL: #8740

cpp/src/io/csv/writer_impl.cu

@@ -285,8 +285,16 @@ void writer::impl::write_chunked_begin(table_view const& table,
                                        const table_metadata* metadata,
                                        rmm::cuda_stream_view stream)
 {
-  if ((metadata != nullptr) && (options_.is_enabled_include_header())) {
-    auto const& column_names = metadata->column_names;
+  if (options_.is_enabled_include_header()) {
+    // need to generate column names if metadata is not provided
+    std::vector<std::string> generated_col_names;
+    if (metadata == nullptr) {
+      generated_col_names.resize(table.num_columns());
+      thrust::tabulate(generated_col_names.begin(), generated_col_names.end(), [](auto idx) {
+        return std::to_string(idx);
+      });
+    }
+    auto const& column_names = (metadata == nullptr) ? generated_col_names : metadata->column_names;
     CUDF_EXPECTS(column_names.size() == static_cast<size_t>(table.num_columns()),
                  "Mismatch between number of column headers and table columns.");
 

cpp/tests/io/csv_test.cpp

@@ -407,7 +407,8 @@ TYPED_TEST(CsvFixedPointWriterTest, SingleColumnNegativeScale)
   auto filepath = temp_env->get_temp_dir() + "FixedPointSingleColumnNegativeScale.csv";
 
   cudf_io::csv_writer_options writer_options =
-    cudf_io::csv_writer_options::builder(cudf_io::sink_info(filepath), input_table);
+    cudf_io::csv_writer_options::builder(cudf_io::sink_info(filepath), input_table)
+      .include_header(false);
 
   cudf_io::write_csv(writer_options);
 
@@ -453,7 +454,8 @@ TYPED_TEST(CsvFixedPointWriterTest, SingleColumnPositiveScale)
   auto filepath = temp_env->get_temp_dir() + "FixedPointSingleColumnPositiveScale.csv";
 
   cudf_io::csv_writer_options writer_options =
-    cudf_io::csv_writer_options::builder(cudf_io::sink_info(filepath), input_table);
+    cudf_io::csv_writer_options::builder(cudf_io::sink_info(filepath), input_table)
+      .include_header(false);
 
   cudf_io::write_csv(writer_options);
 
@@ -2198,4 +2200,32 @@ TEST_F(CsvReaderTest, DtypesMapInvalid)
   EXPECT_THROW(cudf_io::read_csv(in_opts), cudf::logic_error);
 }
 
+TEST_F(CsvReaderTest, CsvDefaultOptionsWriteReadMatch)
+{
+  auto const filepath = temp_env->get_temp_dir() + "issue.csv";
+
+  // make up some kind of dataframe
+  auto int_column = column_wrapper<int32_t>{10, 20, 30};
+  auto str_column = column_wrapper<cudf::string_view>{"abc", "mno", "xyz"};
+  cudf::table_view input_table(std::vector<cudf::column_view>{int_column, str_column});
+
+  // write that dataframe to a csv using default options to some temporary file
+  cudf_io::csv_writer_options writer_options =
+    cudf_io::csv_writer_options::builder(cudf_io::sink_info{filepath}, input_table);
+  cudf_io::write_csv(writer_options);
+
+  // read the temp csv file using default options
+  cudf_io::csv_reader_options read_options =
+    cudf_io::csv_reader_options::builder(cudf_io::source_info{filepath})
+      .dtypes(std::vector<data_type>{dtype<int32_t>(), dtype<cudf::string_view>()});
+
+  cudf_io::table_with_metadata new_table_and_metadata = cudf_io::read_csv(read_options);
+
+  // verify that the tables are identical, or as identical as expected.
+  const auto new_table_view = new_table_and_metadata.tbl->view();
+  CUDF_TEST_EXPECT_TABLES_EQUIVALENT(input_table, new_table_view);
+  EXPECT_EQ(new_table_and_metadata.metadata.column_names[0], "0");
+  EXPECT_EQ(new_table_and_metadata.metadata.column_names[1], "1");
+}
+
 CUDF_TEST_PROGRAM_MAIN()