c# - Task.StartNew Parallel.ForEach 不等待

Question

我有这个代码:

await Task.Factory.StartNew(
    () => Parallel.ForEach(
        urls,
        new ParallelOptions { MaxDegreeOfParallelism = 2 },
        async url =>
        {
           Uri uri = new Uri(url);
           string filename = System.IO.Path.GetFileName(uri.LocalPath);

           using (HttpClient client = new HttpClient())
           using (HttpResponseMessage response = await client.GetAsync(url))
           using (HttpContent content = response.Content)
           {
               // ... Read the string.
               using (var fileStream = new FileStream(config.M_F_P + filename, FileMode.Create, FileAccess.Write))
               {
                   await content.CopyToAsync(fileStream);
               }
           }
        }));

MessageBox.Show("Completed");

它应该处理超过 800 个元素的列表，但它不会等待下载和文件写入完成。 事实上，他开始下载和写入，显示消息，然后在后台继续下载...... 我需要并行和异步下载很多文件，但我必须等待所有文件下载完毕。这段代码有什么问题？

Answer 1

Parallel.ForEach不适用于异步。它需要一个 Action但为了等待异步方法，它需要获得 Func<Task> .

您可以使用 TPL 数据流的 ActionBlock那是在考虑异步的情况下构建的。您给它一个委托(delegate)(异步或非异步)来执行每个项目。您可以配置 block 的并行性(如果需要，还可以配置有限的容量)。然后将您的项目发布到其中:

var block = new ActionBlock<string>(async url => 
{
    Uri uri = new Uri(url);
    string filename = System.IO.Path.GetFileName(uri.LocalPath);

    using (HttpClient client = new HttpClient())
    using (HttpResponseMessage response = await client.GetAsync(url))
    using (HttpContent content = response.Content)
    {
       // ... Read the string.
       using (var fileStream = new FileStream(config.M_F_P + filename, FileMode.Create, FileAccess.Write))
       {
           await content.CopyToAsync(fileStream);
       }
    }
}, new ExecutionDataflowBlockOptions { MaxDegreeOfParallelism = 2 } );

foreach (var url in urls)
{
    block.Post(url);
}

block.Complete();
await block.Completion;
// done