使用C++读取Apache Arrow文件时遇未定义引用错误求助

错误信息 #

链接阶段出现以下未定义引用错误：

read_arrow.cpp:22: undefined reference to `__imp__ZN5arrow19default_memory_poolEv'

代码实现 #

#include <iostream>
#include <memory>
#include <string>

#include <arrow/io/api.h>
#include <arrow/ipc/api.h>
#include <arrow/api.h>
#include <arrow/csv/api.h>
#include <parquet/arrow/reader.h>
#include <parquet/arrow/writer.h>
#include <arrow/status.h>

using namespace arrow;

arrow::Status GenInitialFile()
{

  std::shared_ptr<arrow::io::ReadableFile> file;
  const std::string file_path = "userdata_to_arrow.arrow";
  ARROW_ASSIGN_OR_RAISE(file, arrow::io::ReadableFile::Open(file_path, arrow::default_memory_pool()));

  std::shared_ptr<arrow::ipc::RecordBatchFileReader> reader;
  ARROW_ASSIGN_OR_RAISE(auto ipc_reader, arrow::ipc::RecordBatchFileReader::Open(file));

  std::shared_ptr<arrow::RecordBatch> record_batch;
  arrow::Status status;
  while (true)
  {
    std::shared_ptr<arrow::RecordBatch> rbatch;
    ARROW_ASSIGN_OR_RAISE(rbatch, ipc_reader->ReadRecordBatch(0));
    std::cout << "Read record batch with " << record_batch->num_rows() << " rows." << std::endl;
  }

  return arrow::Status::OK();
}

int main()
{
  
  arrow::Status st = GenInitialFile();

  if (!st.ok())
  {
    std::cerr << st << std::endl;
    return 1;
  }

  return 0;
}

当前c_cpp_properties.json配置 #

{
    "configurations": [
        {
            "name": "Win32",
            "includePath": [
                "C:\\msys64\\mingw64\\**",
                "${vcpkgRoot}/x86-windows/include",
                "${default}",
                "C:\\msys64\\mingw64\\include\\arrow",
                "C:\\msys64\\mingw64\\include\\arrow-flight-glib",
                "C:\\msys64\\mingw64\\bin",
                "C:\\msys64\\mingw64\\include\\arrow-flight-sql-glib",
                "C:\\msys64\\mingw64\\include\\arrow-flight-glib",
                "C:\\msys64\\mingw64\\include\\arrow-glib",
                "C:\\msys64\\mingw64\\include\\arrow-dataset-glib",
                "C:\\msys64\\mingw64\\include\\grpc",
                "C:\\msys64\\mingw64\\include\\grpc++"
            ],
            "defines": [
                "_DEBUG",
                "UNICODE",
                "_UNICODE"
            ],
            "windowsSdkVersion": "10.0.19041.0",
            "cStandard": "c17",
            "cppStandard": "c++17",
            "intelliSenseMode": "${default}",
            "compilerPath": "C:\\msys64\\mingw64\\bin\\g++.exe"
        }
    ],
    "version": 4
}

问题分析与解决 #

这个错误属于链接阶段错误：编译时能找到头文件（所以没有编译错误），但链接器找不到arrow::default_memory_pool()对应的库实现。你的c_cpp_properties.json仅配置了IntelliSense的头文件路径，并未设置编译链接所需的库链接参数。

解决步骤 #

1. 配置tasks.json添加链接参数
   在VS Code中创建或修改.vscode/tasks.json，添加Arrow相关库的链接选项：

   {
     "version": "2.0.0",
     "tasks": [
       {
         "type": "cppbuild",
         "label": "C/C++: g++.exe 生成活动文件",
         "command": "C:\\msys64\\mingw64\\bin\\g++.exe",
         "args": [
           "-fdiagnostics-color=always",
           "-g",
           "${file}",
           "-o",
           "${fileDirname}\\${fileBasenameNoExtension}.exe",
           // 链接Arrow核心库、IPC库和Parquet库
           "-larrow",
           "-larrow_ipc",
           "-lparquet",
           // 如果库不在默认搜索路径，添加库目录
           "-LC:\\msys64\\mingw64\\lib"
         ],
         "options": {
           "cwd": "${fileDirname}"
         },
         "problemMatcher": ["$gcc"],
         "group": {
           "kind": "build",
           "isDefault": true
         },
         "detail": "编译器: C:\\msys64\\mingw64\\bin\\g++.exe"
       }
     ]
   }
   

2. 验证库兼容性
   确保你的Apache Arrow库是用Mingw编译的——Windows下MSVC和Mingw生成的库无法交叉链接，你当前使用Mingw的g++编译器，必须对应使用Mingw编译的Arrow库。

3. 修复代码中的潜在问题
   你的代码存在两处问题：

   • 循环中record_batch未赋值，应使用读取到的rbatch；
   • 死循环会重复读取第0个batch，需改为遍历所有batch：

     int batch_idx = 0;
     while (true) {
       std::shared_ptr<arrow::RecordBatch> rbatch;
       auto status = ipc_reader->ReadRecordBatch(batch_idx, &rbatch);
       if (!status.ok()) {
         if (status.IsEndOfFile()) {
           break; // 读取完毕
         }
         return status;
       }
       std::cout << "Read record batch " << batch_idx << " with " << rbatch->num_rows() << " rows." << std::endl;
       batch_idx++;
     }
     

内容的提问来源于stack exchange，提问作者Aswath