这里要读一个xml文件，并把数据写入csv。

python代码

python 3.8

def read_xml3():
    doc = etree.parse("D:/tmp/bbc_text/bbc-text.xml")
    root=doc.getroot()
    news_items =[]
    with open("D:/tmp/bbc_text/bbc_text_python.csv", "w", encoding="utf-8") as f:
        f.write("category,text\n")
        for news_item_xml in root.getchildren():
            f.write(news_item_xml.attrib['category'])
            f.write(",")
            f.write(news_item_xml.text.replace('"','""'))
            f.write("\n")
        f.flush()
        f.close()

C#代码

    public class BbcNewsItem
    {
   
   
        public string Category {
   
    get; set; }
        public string Text {
   
    get; set; }
    }

    public static class XmlReader
    {
   
   
        public static void Read()
        {
   
   
            XmlDocument xmlDoc = new XmlDocument();
            xmlDoc.Load("D:/tmp/bbc_text/bbc-text.xml");

            List<BbcNewsItem> newsItems = new List<BbcNewsItem>();
            foreach (XmlNode node in xmlDoc.DocumentElement.ChildNodes)
            {
   
   
                XmlElement e = (XmlElement)node;
                BbcNewsItem newsItem = new BbcNewsItem();
                newsItem.Category = e.GetAttribute("category");
                newsItem.Text = e.InnerText;
                newsItems.Add(newsItem);
            }

            using (var writer = new StreamWriter("D:/tmp/bbc_text/bbc_text_cshape.csv"))
            {
   
   
                using (var csvWriter = new CsvWriter(writer, CultureInfo.InvariantCulture))
                {
   
   
                    csvWriter.Configuration.HasHeaderRecord = true;
                    csvWriter.Configuration.AutoMap<BbcNewsItem>();
                    csvWriter.WriteRecords(newsItems);
                    writer.Flush();
                    writer.Close();
                }
            }
        }
    }

测试后所用时间为
235.8ms ± 19.8

不过，我后来发现，我用的是.net core 3.1。因为.net core是跨平台的，所以，并不是windows的机器码。应该是MSIL
于是改成了.net framework 4.5。

57.6 ms ± 10.8

C++代码

C++ 14

#include <iostream>
#include <fstream>
#include <chrono>
#include "pugixml.hpp"

using namespace std;

void findAndReplaceAll(std::string& data, std::string toSearch, std::string replaceStr)
{
   
   
    // Get the first occurrence
    size_t pos = data.find(toSearch);
    // Repeat till end is reached
    while (pos != std::string::npos)
    {
   
   
        // Replace this occurrence of Sub String
        data.replace(pos, toSearch.size(), replaceStr);
        // Get the next occurrence from the current position
        pos = data.find(toSearch, pos + replaceStr.size());
    }
}

int main()
{
   
   
    auto start = chrono::system_clock::now();
    std::cout << "Hello World!\n";

    pugi::xml_document doc;
    pugi::xml_parse_result result = doc.load_file("D:/tmp/bbc_text/bbc-text.xml");
    if (!result)
        return -1;

    ofstream csv_file;
    csv_file.open("D:/tmp/bbc_text/bbc_text_cpp.csv");
    csv_file << "category,text" << endl;
    auto children = doc.child("bbc").children();
    for (auto child : children) {
   
   
        string category = child.attribute("category").value();
        string text = child.text().as_string();
        findAndReplaceAll(text, "\"", "\"\"");
        text = "\"" + text + "\"";
        csv_file << category << "," << text << endl;
    }

    csv_file.flush();
    csv_file.close();

    auto end = chrono::system_clock::now();
    std::chrono::duration<double> elapsed_seconds = end - start;
    auto millis = std::chrono::duration_cast<std::chrono::milliseconds>(elapsed_seconds).count();
    std::cout << millis << "\n";
    string c;
    std::cin >>c;
}