[ARROW-4409] Add convenience to parse JSON from file

kszucs · Feb 5, 2019 · 3e185cb · 3e185cb
1 parent 111b3e6
commit 3e185cb
Show file tree

Hide file tree

Showing 3 changed files with 29 additions and 9 deletions.
diff --git a/cpp/src/arrow/flight/test-integration-client.cc b/cpp/src/arrow/flight/test-integration-client.cc
@@ -51,17 +51,11 @@ int main(int argc, char** argv) {
 
   // 1. Put the data to the server.
   std::unique_ptr<arrow::ipc::internal::json::JsonReader> reader;
-  std::shared_ptr<arrow::io::ReadableFile> in_file;
   std::cout << "Opening JSON file '" << FLAGS_path << "'" << std::endl;
+  std::shared_ptr<arrow::io::ReadableFile> in_file;
   ABORT_NOT_OK(arrow::io::ReadableFile::Open(FLAGS_path, &in_file));
-
-  int64_t file_size = 0;
-  ABORT_NOT_OK(in_file->GetSize(&file_size));
-
-  std::shared_ptr<arrow::Buffer> json_buffer;
-  ABORT_NOT_OK(in_file->Read(file_size, &json_buffer));
-
-  ABORT_NOT_OK(arrow::ipc::internal::json::JsonReader::Open(json_buffer, &reader));
+  ABORT_NOT_OK(arrow::ipc::internal::json::JsonReader::Open(arrow::default_memory_pool(),
+                                                            in_file, &reader));
 
   std::unique_ptr<arrow::ipc::RecordBatchWriter> write_stream;
   ABORT_NOT_OK(client->DoPut(descr, reader->schema(), &write_stream));

diff --git a/cpp/src/arrow/ipc/json.cc b/cpp/src/arrow/ipc/json.cc
@@ -22,6 +22,7 @@
 #include <string>
 
 #include "arrow/buffer.h"
+#include "arrow/io/file.h"
 #include "arrow/ipc/json-internal.h"
 #include "arrow/memory_pool.h"
 #include "arrow/record_batch.h"
@@ -157,6 +158,18 @@ Status JsonReader::Open(MemoryPool* pool, const std::shared_ptr<Buffer>& data,
   return (*reader)->impl_->ParseAndReadSchema();
 }
 
+Status JsonReader::Open(MemoryPool* pool,
+                        const std::shared_ptr<io::ReadableFile>& in_file,
+                        std::unique_ptr<JsonReader>* reader) {
+  int64_t file_size = 0;
+  RETURN_NOT_OK(in_file->GetSize(&file_size));
+
+  std::shared_ptr<arrow::Buffer> json_buffer;
+  RETURN_NOT_OK(in_file->Read(file_size, &json_buffer));
+
+  return Open(pool, json_buffer, reader);
+}
+
 std::shared_ptr<Schema> JsonReader::schema() const { return impl_->schema(); }
 
 int JsonReader::num_record_batches() const { return impl_->num_record_batches(); }

diff --git a/cpp/src/arrow/ipc/json.h b/cpp/src/arrow/ipc/json.h
@@ -33,6 +33,10 @@ class MemoryPool;
 class RecordBatch;
 class Schema;
 
+namespace io {
+class ReadableFile;
+}  // namespace io
+
 namespace ipc {
 namespace internal {
 namespace json {
@@ -95,6 +99,15 @@ class ARROW_EXPORT JsonReader {
   static Status Open(const std::shared_ptr<Buffer>& data,
                      std::unique_ptr<JsonReader>* reader);
 
+  /// \brief Create a new JSON reader from a file
+  ///
+  /// \param[in] pool a MemoryPool to use for buffer allocations
+  /// \param[in] in_file a ReadableFile containing JSON data
+  /// \param[out] reader the returned reader object
+  /// \return Status
+  static Status Open(MemoryPool* pool, const std::shared_ptr<io::ReadableFile>& in_file,
+                     std::unique_ptr<JsonReader>* reader);
+
   /// \brief Return the schema read from the JSON
   std::shared_ptr<Schema> schema() const;