Implement JSON writer

Author: stenlarsson

cpp/src/arrow/CMakeLists.txt

@@ -1000,7 +1000,8 @@ if(ARROW_JSON)
                            json/object_parser.cc
                            json/object_writer.cc
                            json/parser.cc
-                           json/reader.cc)
+                           json/reader.cc
+                           json/writer.cc)
   foreach(ARROW_JSON_TARGET ${ARROW_JSON_TARGETS})
     target_link_libraries(${ARROW_JSON_TARGET} PRIVATE RapidJSON)
   endforeach()

cpp/src/arrow/json/CMakeLists.txt

@@ -23,6 +23,7 @@ add_arrow_test(test
                from_string_test.cc
                parser_test.cc
                reader_test.cc
+               writer_test.cc
                PREFIX
                "arrow-json"
                EXTRA_LINK_LIBS
@@ -33,6 +34,12 @@ add_arrow_benchmark(parser_benchmark
                     "arrow-json"
                     EXTRA_LINK_LIBS
                     RapidJSON)
+
+add_arrow_benchmark(writer_benchmark
+                    PREFIX
+                    "arrow-json"
+                    EXTRA_LINK_LIBS
+                    RapidJSON)
 arrow_install_all_headers("arrow/json")
 
 # pkg-config support

cpp/src/arrow/json/meson.build

@@ -24,6 +24,7 @@ exc = executable(
         'from_string_test.cc',
         'parser_test.cc',
         'reader_test.cc',
+        'writer_test.cc',
     ],
     dependencies: [arrow_test_dep, rapidjson_dep],
 )
@@ -36,6 +37,13 @@ exc = executable(
 )
 benchmark('arrow-json-parser-benchmark', exc)
 
+exc = executable(
+    'arrow-json-writer-benchmark',
+    sources: ['writer_benchmark.cc'],
+    dependencies: [arrow_benchmark_dep, rapidjson_dep],
+)
+benchmark('arrow-json-writer-benchmark', exc)
+
 install_headers(
     [
         'api.h',
@@ -51,6 +59,7 @@ install_headers(
         'reader.h',
         'test_common.h',
         'type_fwd.h',
+        'writer.h',
     ],
     subdir: 'arrow/json',
 )

cpp/src/arrow/json/options.cc

@@ -24,5 +24,14 @@ ParseOptions ParseOptions::Defaults() { return ParseOptions(); }
 
 ReadOptions ReadOptions::Defaults() { return ReadOptions(); }
 
+WriteOptions WriteOptions::Defaults() { return WriteOptions(); }
+
+Status WriteOptions::Validate() const {
+  if (ARROW_PREDICT_FALSE(batch_size < 1)) {
+    return Status::Invalid("WriteOptions: batch_size must be at least 1: ", batch_size);
+  }
+  return Status::OK();
+}
+
 }  // namespace json
 }  // namespace arrow

cpp/src/arrow/json/options.h

@@ -19,8 +19,10 @@
 
 #include <cstdint>
 #include <memory>
+#include <string>
 
 #include "arrow/json/type_fwd.h"
+#include "arrow/status.h"
 #include "arrow/util/visibility.h"
 
 namespace arrow {
@@ -70,5 +72,25 @@ struct ARROW_EXPORT ReadOptions {
   static ReadOptions Defaults();
 };
 
+struct ARROW_EXPORT WriteOptions {
+  /// \brief Maximum number of rows processed at a time
+  ///
+  /// The JSON writer converts and writes data in batches of N rows.
+  /// This number can impact performance.
+  int32_t batch_size = 1024;
+
+  /// \brief Whether to emit null values in the JSON output
+  ///
+  /// If true, null values are included as JSON null.
+  /// If false, null values are omitted from the output entirely.
+  bool emit_null = false;
+
+  /// Create write options with default values
+  static WriteOptions Defaults();
+
+  /// \brief Test that all set options are valid
+  Status Validate() const;
+};
+
 }  // namespace json
 }  // namespace arrow