This is an automated email from the ASF dual-hosted git repository.
pitrou pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new 5cafd7caf7 GH-50706: [C++] Migrate extension type serialization to
JsonWriter (#50708)
5cafd7caf7 is described below
commit 5cafd7caf72bcdb7a56fa7c3196b24c8091cb050
Author: Aaditya Srinivasan <[email protected]>
AuthorDate: Wed Jul 29 18:56:24 2026 +0530
GH-50706: [C++] Migrate extension type serialization to JsonWriter (#50708)
### Rationale for this change
This PR continues the simdjson migration by replacing RapidJSON's `Writer`
API with `JsonWriter` in extension type serialization.
### What changes are included in this PR?
* Replace RapidJSON writer usage with `JsonWriter` in:
* `FixedShapeTensorType::Serialize()`
* `VariableShapeTensorType::Serialize()`
* `OpaqueType::Serialize()`
* Remove `rapidjson::Writer` usage from the migrated serializers.
### Are these changes tested?
Yes.
### Are there any user-facing changes?
No.
Closes: #50706
* GitHub Issue: #50706
Authored-by: Aaditya Srinivasan <[email protected]>
Signed-off-by: Antoine Pitrou <[email protected]>
---
cpp/src/arrow/extension/fixed_shape_tensor.cc | 39 +++++++++++----------
cpp/src/arrow/extension/opaque.cc | 27 +++++++--------
cpp/src/arrow/extension/variable_shape_tensor.cc | 43 +++++++++++++-----------
3 files changed, 57 insertions(+), 52 deletions(-)
diff --git a/cpp/src/arrow/extension/fixed_shape_tensor.cc
b/cpp/src/arrow/extension/fixed_shape_tensor.cc
index 5446169887..697bab92b3 100644
--- a/cpp/src/arrow/extension/fixed_shape_tensor.cc
+++ b/cpp/src/arrow/extension/fixed_shape_tensor.cc
@@ -25,6 +25,7 @@
#include "arrow/array/array_nested.h"
#include "arrow/array/array_primitive.h"
+#include "arrow/json/json_writer_internal.h"
#include "arrow/json/rapidjson_defs.h" // IWYU pragma: keep
#include "arrow/tensor.h"
#include "arrow/util/logging_internal.h"
@@ -33,9 +34,9 @@
#include "arrow/util/string.h"
#include <rapidjson/document.h>
-#include <rapidjson/writer.h>
namespace rj = arrow::rapidjson;
+using ::arrow::json::JsonWriter;
namespace arrow::extension {
@@ -72,36 +73,38 @@ std::string FixedShapeTensorType::ToString(bool
show_metadata) const {
}
std::string FixedShapeTensorType::Serialize() const {
- rj::Document document;
- document.SetObject();
- rj::Document::AllocatorType& allocator = document.GetAllocator();
+ JsonWriter writer;
+
+ writer.StartObject();
- rj::Value shape(rj::kArrayType);
+ writer.Key("shape");
+ writer.StartArray();
for (auto v : shape_) {
- shape.PushBack(v, allocator);
+ writer.Int64(v);
}
- document.AddMember(rj::Value("shape", allocator), shape, allocator);
+ writer.EndArray();
if (!permutation_.empty()) {
- rj::Value permutation(rj::kArrayType);
+ writer.Key("permutation");
+ writer.StartArray();
for (auto v : permutation_) {
- permutation.PushBack(v, allocator);
+ writer.Int64(v);
}
- document.AddMember(rj::Value("permutation", allocator), permutation,
allocator);
+ writer.EndArray();
}
if (!dim_names_.empty()) {
- rj::Value dim_names(rj::kArrayType);
- for (const std::string& v : dim_names_) {
- dim_names.PushBack(rj::Value{}.SetString(v.c_str(), allocator),
allocator);
+ writer.Key("dim_names");
+ writer.StartArray();
+ for (const auto& v : dim_names_) {
+ writer.String(v);
}
- document.AddMember(rj::Value("dim_names", allocator), dim_names,
allocator);
+ writer.EndArray();
}
- rj::StringBuffer buffer;
- rj::Writer<rj::StringBuffer> writer(buffer);
- document.Accept(writer);
- return buffer.GetString();
+ writer.EndObject();
+
+ return std::string(writer.GetString());
}
Result<std::shared_ptr<DataType>> FixedShapeTensorType::Deserialize(
diff --git a/cpp/src/arrow/extension/opaque.cc
b/cpp/src/arrow/extension/opaque.cc
index e206007ebe..5fe904fce6 100644
--- a/cpp/src/arrow/extension/opaque.cc
+++ b/cpp/src/arrow/extension/opaque.cc
@@ -19,12 +19,14 @@
#include <sstream>
+#include "arrow/json/json_writer_internal.h"
#include "arrow/json/rapidjson_defs.h" // IWYU pragma: keep
#include "arrow/util/logging_internal.h"
#include <rapidjson/document.h>
#include <rapidjson/error/en.h>
-#include <rapidjson/writer.h>
+
+using ::arrow::json::JsonWriter;
namespace arrow::extension {
@@ -46,19 +48,16 @@ bool OpaqueType::ExtensionEquals(const ExtensionType&
other) const {
}
std::string OpaqueType::Serialize() const {
- rapidjson::Document document;
- document.SetObject();
- rapidjson::Document::AllocatorType& allocator = document.GetAllocator();
-
- rapidjson::Value type_name(rapidjson::StringRef(type_name_));
- document.AddMember(rapidjson::Value("type_name", allocator), type_name,
allocator);
- rapidjson::Value vendor_name(rapidjson::StringRef(vendor_name_));
- document.AddMember(rapidjson::Value("vendor_name", allocator), vendor_name,
allocator);
-
- rapidjson::StringBuffer buffer;
- rapidjson::Writer<rapidjson::StringBuffer> writer(buffer);
- document.Accept(writer);
- return buffer.GetString();
+ JsonWriter writer;
+
+ writer.StartObject();
+
+ writer.StringField("type_name", type_name_);
+ writer.StringField("vendor_name", vendor_name_);
+
+ writer.EndObject();
+
+ return std::string(writer.GetString());
}
Result<std::shared_ptr<DataType>> OpaqueType::Deserialize(
diff --git a/cpp/src/arrow/extension/variable_shape_tensor.cc
b/cpp/src/arrow/extension/variable_shape_tensor.cc
index b1b12583d7..a67bd6dea8 100644
--- a/cpp/src/arrow/extension/variable_shape_tensor.cc
+++ b/cpp/src/arrow/extension/variable_shape_tensor.cc
@@ -21,6 +21,7 @@
#include "arrow/extension/variable_shape_tensor.h"
#include "arrow/array/array_primitive.h"
+#include "arrow/json/json_writer_internal.h"
#include "arrow/json/rapidjson_defs.h" // IWYU pragma: keep
#include "arrow/scalar.h"
#include "arrow/tensor.h"
@@ -30,9 +31,9 @@
#include "arrow/util/string.h"
#include <rapidjson/document.h>
-#include <rapidjson/writer.h>
namespace rj = arrow::rapidjson;
+using ::arrow::json::JsonWriter;
namespace arrow::extension {
@@ -82,42 +83,44 @@ std::string VariableShapeTensorType::ToString(bool
show_metadata) const {
}
std::string VariableShapeTensorType::Serialize() const {
- rj::Document document;
- document.SetObject();
- rj::Document::AllocatorType& allocator = document.GetAllocator();
+ JsonWriter writer;
+
+ writer.StartObject();
if (!permutation_.empty()) {
- rj::Value permutation(rj::kArrayType);
+ writer.Key("permutation");
+ writer.StartArray();
for (auto v : permutation_) {
- permutation.PushBack(v, allocator);
+ writer.Int64(v);
}
- document.AddMember(rj::Value("permutation", allocator), permutation,
allocator);
+ writer.EndArray();
}
if (!dim_names_.empty()) {
- rj::Value dim_names(rj::kArrayType);
- for (const std::string& v : dim_names_) {
- dim_names.PushBack(rj::Value{}.SetString(v.c_str(), allocator),
allocator);
+ writer.Key("dim_names");
+ writer.StartArray();
+ for (const auto& v : dim_names_) {
+ writer.String(v);
}
- document.AddMember(rj::Value("dim_names", allocator), dim_names,
allocator);
+ writer.EndArray();
}
if (!uniform_shape_.empty()) {
- rj::Value uniform_shape(rj::kArrayType);
- for (auto v : uniform_shape_) {
+ writer.Key("uniform_shape");
+ writer.StartArray();
+ for (const auto& v : uniform_shape_) {
if (v.has_value()) {
- uniform_shape.PushBack(v.value(), allocator);
+ writer.Int64(*v);
} else {
- uniform_shape.PushBack(rj::Value{}.SetNull(), allocator);
+ writer.Null();
}
}
- document.AddMember(rj::Value("uniform_shape", allocator), uniform_shape,
allocator);
+ writer.EndArray();
}
- rj::StringBuffer buffer;
- rj::Writer<rj::StringBuffer> writer(buffer);
- document.Accept(writer);
- return buffer.GetString();
+ writer.EndObject();
+
+ return std::string(writer.GetString());
}
Result<std::shared_ptr<DataType>> VariableShapeTensorType::Deserialize(
