如何迭代所有行并有效地在 EF Core 中应用更新？

Question

我需要遍历与 Songs 具有一对多关系的整个 Artist table。艺术家和歌曲列表的详细信息存储在一个文件中，此路径保存在 Artist table、DataFilePath.

下面的代码是我对此的尝试，但它给出了以下异常： System.InvalidOperationException: Connection is busy

// Artist to Songs is a 1 to many relationship
foreach(var artist in db.Artist.Include(x => x.Songs)) 
{
    try
    {
        // Load details for artist and songs from artist.DataFilePath
        var fileData = LoadArtistFile(artist.DataFilePath);

        // UpdateArtists() will
        //  - update each property for the Artists entity
        //  - add any new Songs 
        //  - update each Song property for existing songs
        UpdateArtist(artist, fileData);
        
        db.SaveChanges();
    } catch (Exception err)
    {
        _logger.LogError(err.Message);
    }
}

2 个问题：

如何修复上面的代码以避免连接忙问题？（我尝试了很多其他变体，包括延迟加载和显式加载）。
如果艺术家 table 有数百万行，因此歌曲 table 有数百万行 - 这仍然是更新两者的最有效方法吗 tables？（留在 EF Core 中）

更新

我采纳了 David 的回答并修改了使用批处理的方法，因为我无法一次将所有行加载到内存中...

int batchSize = 50; // batch of 50 gave best performance (vs 10 and 100)
int currentBatch = 0;
// get the max id from the database for a stopping point (probably a better way to do this)
int maxId = GetMaxArtistID();
bool done = false;
while (!done)
{
    using (var db = new Context())
        {
            var artists = db.Artists
                .OrderBy(x => x.Id)
                .Skip(currentBatch++ * batchSize)
                .Take(batchSize)
                .Include(x => x.Songs)
                .ToList();
            var ids = String.Join(", ", artists.Select(x => x.Id));
            Console.WriteLine($"Working on batch:{currentBatch} ids: {ids}");
            foreach (var artist in artists)
            {
                var fileData = LoadArtistFile(artist.DataFilePath)
                UpdateArtist(artist, fileData);
                if (artist.Id >= maxId)
                    done = true;
            }
            db.SaveChanges();
        }
    }
}

Answer 1

How do I fix the code above to avoid the Connection is busy problem?

在迭代之前将数据加载到内存中。例如

foreach(var artist in db.Artist.Include(x => x.Songs).ToList())

If the artists table were millions of rows and therefore songs table were 10's of millions of rows - is this still the most effective way to update the both tables? (staying within EF Core)

EF 本身不做服务器端数据修改，所以这确实是唯一的方法。但在某种程度上，您希望不使用 EF，并将您的数据文件加载到 table 并在服务器端执行更新。

如何迭代所有行并有效地在 EF Core 中应用更新？

How does one iterate over all rows and apply updates in EF Core effectively?

entity-framework

one-to-many

entity-framework-core

.net-core

更新