GH-50859: [C++][Parquet] Move JsonWriter to simdjson utilities

cpp/cmake_modules/DefineOptions.cmake

@@ -599,8 +599,7 @@ takes precedence over ccache if a storage backend is configured" ON)
                 "Build support for encryption. Fail if OpenSSL is not found"
                 OFF
                 DEPENDS
-                ARROW_FILESYSTEM
-                ARROW_JSON)
+                ARROW_FILESYSTEM)
 
   #----------------------------------------------------------------------
   set_option_category("Gandiva")

cpp/cmake_modules/ThirdpartyToolchain.cmake

@@ -383,7 +383,6 @@ if(ARROW_WITH_OPENTELEMETRY)
 endif()
 
 if(ARROW_PARQUET)
-  set(ARROW_WITH_RAPIDJSON ON)
   set(ARROW_WITH_SIMDJSON ON)
   set(ARROW_WITH_THRIFT ON)
 endif()
@@ -411,11 +410,11 @@ if(ARROW_AZURE)
   set(ARROW_WITH_AZURE_SDK ON)
 endif()
 
-if(ARROW_JSON OR ARROW_FLIGHT_SQL_ODBC)
+if(ARROW_JSON)
   set(ARROW_WITH_RAPIDJSON ON)
 endif()
 
-if(ARROW_JSON)
+if(ARROW_JSON OR ARROW_FLIGHT_SQL_ODBC)
   set(ARROW_WITH_SIMDJSON ON)
 endif()
 

cpp/meson.build

@@ -104,6 +104,7 @@ needs_testing = (get_option('testing').enabled()
     or needs_integration
 )
 needs_json = get_option('json').enabled() or needs_testing
+needs_simdjson = needs_json or needs_parquet
 needs_brotli = get_option('brotli').enabled() or needs_fuzzing
 needs_bz2 = get_option('bz2').enabled()
 needs_lz4 = get_option('lz4').enabled()

cpp/src/arrow/CMakeLists.txt

@@ -644,6 +644,13 @@ endif()
 if(ARROW_WITH_OPENTELEMETRY)
   list(APPEND ARROW_UTIL_SRCS util/tracing_internal.cc)
 endif()
+if(ARROW_WITH_SIMDJSON)
+  list(APPEND
+       ARROW_UTIL_SRCS
+       json/object_parser.cc
+       util/json_writer_internal.cc
+       util/simdjson_internal.cc)
+endif()
 if(ARROW_WITH_SNAPPY)
   list(APPEND ARROW_UTIL_SRCS util/compression_snappy.cc)
 endif()
@@ -664,6 +671,12 @@ foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
   target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE ${ARROW_XSIMD})
 endforeach()
 
+if(ARROW_WITH_SIMDJSON)
+  foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
+    target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE arrow::simdjson)
+  endforeach()
+endif()
+
 if(ARROW_USE_BOOST)
   foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
     target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE Boost::headers)
@@ -1088,8 +1101,6 @@ if(ARROW_JSON)
                            json/chunker.cc
                            json/converter.cc
                            json/from_string.cc
-                           json/json_writer_internal.cc
-                           json/object_parser.cc
                            json/parser.cc
                            json/reader.cc)
   foreach(ARROW_JSON_TARGET ${ARROW_JSON_TARGETS})

cpp/src/arrow/dataset/file_json_test.cc

@@ -20,10 +20,10 @@
 #include "arrow/dataset/plan.h"
 #include "arrow/dataset/test_util_internal.h"
 #include "arrow/filesystem/mockfs.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/json/parser.h"
 #include "arrow/testing/gtest_util.h"
 #include "arrow/testing/util.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/logging_internal.h"
 
 namespace arrow {

cpp/src/arrow/extension/fixed_shape_tensor.cc

@@ -27,8 +27,8 @@
 
 #include "arrow/array/array_nested.h"
 #include "arrow/array/array_primitive.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/tensor.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/logging_internal.h"
 #include "arrow/util/print_internal.h"
 #include "arrow/util/simdjson_internal.h"

cpp/src/arrow/extension/opaque.cc

@@ -19,7 +19,7 @@
 
 #include <sstream>
 
-#include "arrow/json/json_writer_internal.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/logging_internal.h"
 #include "arrow/util/simdjson_internal.h"
 

cpp/src/arrow/extension/variable_shape_tensor.cc

@@ -23,9 +23,9 @@
 #include "arrow/extension/variable_shape_tensor.h"
 
 #include "arrow/array/array_primitive.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/scalar.h"
 #include "arrow/tensor.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/logging_internal.h"
 #include "arrow/util/print_internal.h"
 #include "arrow/util/simdjson_internal.h"

cpp/src/arrow/flight/sql/odbc/odbc_impl/json_converter.cc

@@ -20,8 +20,8 @@
 #include <boost/beast/core/detail/base64.hpp>
 #include "arrow/builder.h"
 #include "arrow/flight/sql/odbc/odbc_impl/util.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/scalar.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/visitor.h"
 
 using boost::beast::detail::base64::encode;

cpp/src/arrow/integration/json_integration.cc

@@ -29,11 +29,11 @@
 #include "arrow/integration/json_internal.h"
 #include "arrow/io/file.h"
 #include "arrow/ipc/dictionary.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/record_batch.h"
 #include "arrow/result.h"
 #include "arrow/status.h"
 #include "arrow/type.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/logging_internal.h"
 #include "arrow/util/simdjson_internal.h"
 

cpp/src/arrow/integration/json_integration_test.cc

@@ -38,7 +38,6 @@
 #include "arrow/ipc/reader.h"
 #include "arrow/ipc/test_common.h"
 #include "arrow/ipc/writer.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/pretty_print.h"
 #include "arrow/status.h"
 #include "arrow/testing/builder.h"
@@ -49,6 +48,7 @@
 #include "arrow/type.h"
 #include "arrow/type_fwd.h"
 #include "arrow/util/io_util.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/simdjson_internal.h"
 
 DEFINE_string(arrow, "", "Arrow file name");

cpp/src/arrow/integration/json_internal.cc

@@ -36,7 +36,6 @@
 #include "arrow/array/builder_time.h"
 #include "arrow/extension_type.h"
 #include "arrow/ipc/dictionary.h"
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/record_batch.h"
 #include "arrow/result.h"
 #include "arrow/scalar.h"
@@ -48,6 +47,7 @@
 #include "arrow/util/checked_cast.h"
 #include "arrow/util/decimal.h"
 #include "arrow/util/formatting.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/key_value_metadata.h"
 #include "arrow/util/logging_internal.h"
 #include "arrow/util/range.h"

cpp/src/arrow/json/CMakeLists.txt

@@ -21,7 +21,6 @@ add_arrow_test(test
                chunker_test.cc
                converter_test.cc
                from_string_test.cc
-               json_writer_internal_test.cc
                object_parser_test.cc
                parser_test.cc
                reader_test.cc

cpp/src/arrow/json/meson.build

@@ -22,7 +22,6 @@ exc = executable(
         'chunker_test.cc',
         'converter_test.cc',
         'from_string_test.cc',
-        'json_writer_internal_test.cc',
         'parser_test.cc',
         'reader_test.cc',
     ],

cpp/src/arrow/json/object_parser.h

@@ -29,7 +29,7 @@ namespace json {
 namespace internal {
 
 /// This class is a helper to parse a json object from a string.
-/// It uses rapidjson::Document in implementation.
+/// It uses simdjson in the implementation.
 class ARROW_EXPORT ObjectParser {
  public:
   ObjectParser();

cpp/src/arrow/meson.build

@@ -18,6 +18,22 @@
 dl_dep = dependency('dl')
 threads_dep = dependency('threads')
 
+if needs_simdjson
+    simdjson_dep = dependency('simdjson', allow_fallback: false, required: false)
+
+    if not simdjson_dep.found()
+        cmake = import('cmake')
+        simdjson_opts = cmake.subproject_options()
+
+        simdjson_opts.add_cmake_defines({'SIMDJSON_EXCEPTIONS': 'OFF'})
+
+        simdjson_proj = cmake.subproject('simdjson', options: simdjson_opts)
+        simdjson_dep = simdjson_proj.dependency('simdjson')
+    endif
+else
+    simdjson_dep = disabler()
+endif
+
 arrow_components = {
     'arrow_array': {
         'sources': [
@@ -224,6 +240,15 @@ arrow_util_srcs = [
 
 arrow_util_deps = [threads_dep]
 
+if needs_simdjson
+    arrow_util_srcs += [
+        'json/object_parser.cc',
+        'util/json_writer_internal.cc',
+        'util/simdjson_internal.cc',
+    ]
+    arrow_util_deps += [simdjson_dep]
+endif
+
 if needs_brotli
     arrow_util_srcs += ['util/compression_brotli.cc']
     arrow_util_deps += [dependency('libbrotlidec'), dependency('libbrotlienc')]
@@ -318,22 +343,6 @@ else
     rapidjson_dep = disabler()
 endif
 
-if needs_json or needs_integration
-    simdjson_dep = dependency('simdjson', allow_fallback: false, required: false)
-
-    if not simdjson_dep.found()
-        cmake = import('cmake')
-        simdjson_opts = cmake.subproject_options()
-
-        simdjson_opts.add_cmake_defines({'SIMDJSON_EXCEPTIONS': 'OFF'})
-
-        simdjson_proj = cmake.subproject('simdjson', options: simdjson_opts)
-        simdjson_dep = simdjson_proj.dependency('simdjson')
-    endif
-else
-    simdjson_dep = disabler()
-endif
-
 azure_dep = disabler()
 gcs_dep = disabler()
 s3_dep = disabler()
@@ -530,8 +539,6 @@ if needs_json
                 'json/chunker.cc',
                 'json/converter.cc',
                 'json/from_string.cc',
-                'json/json_writer_internal.cc',
-                'json/object_parser.cc',
                 'json/options.cc',
                 'json/parser.cc',
                 'json/reader.cc',

cpp/src/arrow/util/CMakeLists.txt

@@ -87,6 +87,14 @@ add_arrow_test(utility-test
                EXTRA_LINK_LIBS
                ${ARROW_UTILITY_TEST_LINK_LIBS})
 
+if(ARROW_WITH_SIMDJSON)
+  add_arrow_test(json-writer-internal-test
+                 SOURCES
+                 json_writer_internal_test.cc
+                 EXTRA_LINK_LIBS
+                 arrow::simdjson)
+endif()
+
 add_arrow_test(async-utility-test
                SOURCES
                async_generator_test.cc

cpp/src/arrow/json/json_writer_internal.cc → cpp/src/arrow/util/json_writer_internal.cc

@@ -15,7 +15,7 @@
 // specific language governing permissions and limitations
 // under the License.
 
-#include "arrow/json/json_writer_internal.h"
+#include "arrow/util/json_writer_internal.h"
 #include "arrow/util/simdjson_internal.h"
 
 namespace arrow::json {
@@ -100,21 +100,21 @@ void JsonWriter::Double(double value) {
 }
 
 Status JsonWriter::WriteValue(sj::value value) {
-  return internal::VisitJsonValue(
+  return ::arrow::internal::VisitJsonValue(
       value,
 
       [&](sj::object object) -> Status {
         StartObject();
 
         for (auto field : object) {
-          ARROW_ASSIGN_OR_RAISE(
-              auto key, internal::ResolveSimdjsonResult(field.unescaped_key(),
-                                                        "Failed to get object key"));
+          ARROW_ASSIGN_OR_RAISE(auto key,
+                                ::arrow::internal::ResolveSimdjsonResult(
+                                    field.unescaped_key(), "Failed to get object key"));
 
           Key(key);
 
           ARROW_ASSIGN_OR_RAISE(auto field_value,
-                                internal::ResolveSimdjsonResult(
+                                ::arrow::internal::ResolveSimdjsonResult(
                                     field.value(), "Failed to get object value"));
 
           RETURN_NOT_OK(WriteValue(field_value));
@@ -128,9 +128,9 @@ Status JsonWriter::WriteValue(sj::value value) {
         StartArray();
 
         for (auto element : array) {
-          ARROW_ASSIGN_OR_RAISE(
-              auto element_value,
-              internal::ResolveSimdjsonResult(element, "Failed to iterate JSON array"));
+          ARROW_ASSIGN_OR_RAISE(auto element_value,
+                                ::arrow::internal::ResolveSimdjsonResult(
+                                    element, "Failed to iterate JSON array"));
 
           RETURN_NOT_OK(WriteValue(element_value));
         }
@@ -170,7 +170,7 @@ Status JsonWriter::WriteValue(sj::value value) {
       },
 
       [&](sj::value value) -> Status {
-        ARROW_ASSIGN_OR_RAISE(auto raw_json, internal::ResolveSimdjsonResult(
+        ARROW_ASSIGN_OR_RAISE(auto raw_json, ::arrow::internal::ResolveSimdjsonResult(
                                                  simdjson::to_json_string(value),
                                                  "Failed to get raw JSON"));
         RawValue(raw_json);

cpp/src/arrow/json/json_writer_internal_test.cc → cpp/src/arrow/util/json_writer_internal_test.cc

@@ -17,8 +17,8 @@
 
 #include <gtest/gtest.h>
 
-#include "arrow/json/json_writer_internal.h"
 #include "arrow/testing/gtest_util.h"
+#include "arrow/util/json_writer_internal.h"
 
 namespace sj = simdjson::ondemand;
 

cpp/src/arrow/util/meson.build

@@ -222,6 +222,12 @@ utility_test_srcs = [
     'value_parsing_test.cc',
 ]
 
+utility_test_deps = [arrow_test_dep_no_main]
+if needs_simdjson
+    utility_test_srcs += ['json_writer_internal_test.cc']
+    utility_test_deps += [simdjson_dep]
+endif
+
 if host_machine.system() == 'windows'
     # This manifest enables long file paths on Windows 10+
     # See https://docs.microsoft.com/en-us/windows/win32/fileio/naming-a-file#enable-long-paths-in-windows-10-version-1607-and-later
@@ -235,7 +241,7 @@ endif
 exc = executable(
     'arrow-utility-test',
     sources: utility_test_srcs,
-    dependencies: arrow_test_dep_no_main,
+    dependencies: utility_test_deps,
     implicit_include_directories: false,
 )
 test('arrow-utility-test', exc)