LINQ 不会使用 EF Core 对组产生 COUNT() 查询

Question

我有一个 ASP.NET Core 应用程序，它通过 EF Core 在 SQL Server 2012 上运行。我想要实现的是计算不同组的大小，例如

from b in this._context.Benchmark
group b by b.Device into g
select new {
    Device = g.Key,
    Count = g.Count()
}

问题是整个事情非常慢，原因似乎是SQL语句没有映射到GROUP BY和COUNT(*)，但是根据调试器，应用程序请求所有内容并且在 CPU 上执行计数。我从调试器的事件窗口得到的查询如下：

SELECT [b0].[ID], [b0].[CreateDate], [b0].[Creator], [b0].[Device], [b0].[Environment], [b0].[Machine], [b0].[Name], [b0].[Plugin], [b0].[RunDate]
FROM [Benchmark] AS [b0]
ORDER BY [b0].[Device]

我还可以看到，我用于调试的 IIS Express 的内存消耗对于简单的COUNT() 来说是疯狂的，所以我倾向于相信这是实际的查询。

问题是：我如何改写查询，使其实际上映射到COUNT()？

编辑：我在“真实”EF 和 ctx.Database.Log = Console.Write 上尝试了相同的查询，这会产生预期的 COUNT() 查询，这让我相信这是 EF Core 的问题.

Answer 1

请将 EF Core 升级到 2.1.2 版本，现在支持。

Answer 2

无法翻译 LINQ 表达式“GroupBy([x.Device], [x])” 并将在本地进行评估。

无法翻译 LINQ 表达式“Count()”并将 本地评估。

以上日志来自我后来的测试结果，显示GroupBy&Count()尚不支持将完整的LINQ转换为SQL，然后直接在数据库中执行完整的SQL，现在EF Core必须首先获取所有数据，然后在本地获取 GroupBy 结果。所以，如果你的数据很大，那么性能会很差。

您必须硬编码一个简单的 SQL 才能直接在数据库中执行您的工作，然后性能将恢复正常。

我升级到 EF Core 2.1.2，在MySQL 和MS SQL 中都尝试了以下代码：

var query = _context.Benchmark.GroupBy(x => x.Device, 
              (key, group) => new  //Result selector
              {
                  Device = key,
                  Count = group.Count()
              });
await query.ToListAsync();

结果选择器就在GroupBy 内部，这将确保它是否可以转换为正确的 SQL，不幸的是，它不是。

这里还有一个棘手的问题，避免使用Count：

var query = _context.Benchmark.GroupBy(x => x.Device, 
              (key, group) => new  //Result selector
              {
                  Device = key,
                  Count = group.Select(g => g.Id)//Avoid to use Count
              });
await query.ToListAsync();

C# 代码通过this tool 生成以下日志：

2018-08-24T14:27:26.6737424-04:00 Information 10403 Microsoft.EntityFrameworkCore.Infrastructure
Entity Framework Core 2.1.2-rtm-30932 initialized 'ApplicationDbContext' using provider 'Pomelo.EntityFrameworkCore.MySql' with options: None
2018-08-24T14:27:31.4270317-04:00 Debug 10101 Microsoft.EntityFrameworkCore.Query
Compiling query model: 
'from IGrouping<Device, Benchmark> <generated>_0 in 
    (from Benchmark x in DbSet<Benchmark>
    select [x]).GroupBy([x].Device, [x])
select new <>f__AnonymousType3<Device, int>(
    [<generated>_0].Key, 

        (from Benchmark <generated>_1 in [<generated>_0]
        select [<generated>_1]).Count()
)'
2018-08-24T14:27:31.4319437-04:00 Debug 10104 Microsoft.EntityFrameworkCore.Query
Optimized query model: 
'from IGrouping<Device, Benchmark> <generated>_0 in 
    (from Benchmark x in DbSet<Benchmark>
    join Device x.Device in DbSet<Device>
    on Property([x], "DeviceId") equals Property([x.Device], "Id")
    select [x]).GroupBy([x.Device], [x])
select new <>f__AnonymousType3<Device, int>(
    [<generated>_0].Key, 

        (from Benchmark <generated>_1 in [<generated>_0]
        select [<generated>_1]).Count()
)'

这里是核心 LOGS，显示 GROUPBY 尚不支持：

2018-08-24T14:27:31.4431635-04:00 Warning 20500 Microsoft.EntityFrameworkCore.Query
The LINQ expression 'GroupBy([x.Device], [x])' could not be translated and will be evaluated locally.
2018-08-24T14:27:31.4476637-04:00 Warning 20500 Microsoft.EntityFrameworkCore.Query
The LINQ expression 'Count()' could not be translated and will be evaluated locally.
2018-08-24T14:27:31.4511652-04:00 Warning 20500 Microsoft.EntityFrameworkCore.Query
The LINQ expression 'Count()' could not be translated and will be evaluated locally.

这是之后的休息日志：

2018-08-24T14:27:31.4608060-04:00 Debug 10107 Microsoft.EntityFrameworkCore.Query
(QueryContext queryContext) => IAsyncEnumerable<<>f__AnonymousType3<Device, int>> _InterceptExceptions(
    source: IAsyncEnumerable<<>f__AnonymousType3<Device, int>> _SelectAsync(
        source: IAsyncEnumerable<IGrouping<Device, ValueBuffer>> _GroupBy(
            source: IAsyncEnumerable<TransparentIdentifier<ValueBuffer, Device>> _ShapedQuery(
                queryContext: queryContext, 
                shaperCommandContext: SelectExpression: 
                    SELECT `x.Device`.`Id`, `x.Device`.`Description`, `x.Device`.`IP`, `x.Device`.`Name`, `x.Device`.`Port`
                    FROM `Benchmark` AS `x`
                    INNER JOIN `Device` AS `x.Device` ON `x`.`DeviceId` = `x.Device`.`Id`
                    ORDER BY `x.Device`.`Id`, 
                shaper: TypedCompositeShaper<ValueBufferShaper, ValueBuffer, BufferedOffsetEntityShaper<Device>, Device, TransparentIdentifier<ValueBuffer, Device>>), 
            keySelector: (TransparentIdentifier<ValueBuffer, Device> t0) => t0.Inner, 
            elementSelector: (TransparentIdentifier<ValueBuffer, Device> t0) => t0.Outer), 
        selector: (IGrouping<Device, ValueBuffer> <generated>_0 | CancellationToken ct) => Task<<>f__AnonymousType3<Device, int>> _ExecuteAsync(
            taskFactories: new Func<Task<object>>[]{ () => Task<object> _ToObjectTask(Task<int> Count(
                        source: IAsyncEnumerable<ValueBuffer> _ToAsyncEnumerable(<generated>_0), 
                        cancellationToken: queryContext.CancellationToken)) }, 
            selector: (object[] results) => new <>f__AnonymousType3<Device, int>(
                <generated>_0.Key, 
                (int)results[0]
            ))), 
    contextType: Sequencer.Updater.Data.ApplicationDbContext, 
    logger: DiagnosticsLogger<Query>, 
    queryContext: queryContext)
2018-08-24T14:27:31.4825759-04:00 Debug 20000 Microsoft.EntityFrameworkCore.Database.Connection
Opening connection to database 'TestDB' on server 'localhost,3306'.
2018-08-24T14:27:31.4877302-04:00 Debug 20001 Microsoft.EntityFrameworkCore.Database.Connection
Opened connection to database 'TestDB' on server 'localhost,3306'.
2018-08-24T14:27:31.4901269-04:00 Debug 20100 Microsoft.EntityFrameworkCore.Database.Command
Executing DbCommand [Parameters=[], CommandType='Text', CommandTimeout='30']
SELECT `x.Device`.`Id`, `x.Device`.`Description`, `x.Device`.`IP`, `x.Device`.`Name`, `x.Device`.`Port`
FROM `Benchmark` AS `x`
INNER JOIN `Device` AS `x.Device` ON `x`.`DeviceId` = `x.Device`.`Id`
ORDER BY `x.Device`.`Id`
2018-08-24T14:27:31.4929857-04:00 Information 20101 Microsoft.EntityFrameworkCore.Database.Command
Executed DbCommand (3ms) [Parameters=[], CommandType='Text', CommandTimeout='30']
SELECT `x.Device`.`Id`, `x.Device`.`Description`, `x.Device`.`IP`, `x.Device`.`Name`, `x.Device`.`Port`
FROM `Benchmark` AS `x`
INNER JOIN `Device` AS `x.Device` ON `x`.`DeviceId` = `x.Device`.`Id`
ORDER BY `x.Device`.`Id`
2018-08-24T14:27:31.5231128-04:00 Debug 20300 Microsoft.EntityFrameworkCore.Database.Command
A data reader was disposed.
2018-08-24T14:27:31.5270399-04:00 Debug 20002 Microsoft.EntityFrameworkCore.Database.Connection
Closing connection to database 'TestDB' on server 'localhost,3306'.
2018-08-24T14:27:31.5303748-04:00 Debug 20003 Microsoft.EntityFrameworkCore.Database.Connection
Closed connection to database 'TestDB' on server 'localhost,3306'.

这是我的环境：

PM> dotnet --info
.NET Core SDK (reflecting any global.json):
 Version:   2.1.401
 Commit:    91b1c13032

Runtime Environment:
 OS Name:     Windows
 OS Version:  10.0.16299
 OS Platform: Windows
 RID:         win10-x64
 Base Path:   C:\Program Files\dotnet\sdk\2.1.401\

Host (useful for support):
  Version: 2.1.3
  Commit:  124038c13e

.NET Core SDKs installed:
  2.1.202 [C:\Program Files\dotnet\sdk]
  2.1.302 [C:\Program Files\dotnet\sdk]
  2.1.400 [C:\Program Files\dotnet\sdk]
  2.1.401 [C:\Program Files\dotnet\sdk]

.NET Core runtimes installed:
  Microsoft.AspNetCore.All 2.1.2 [C:\Program Files\dotnet\shared\Microsoft.AspNetCore.All]
  Microsoft.AspNetCore.All 2.1.3 [C:\Program Files\dotnet\shared\Microsoft.AspNetCore.All]
  Microsoft.AspNetCore.App 2.1.2 [C:\Program Files\dotnet\shared\Microsoft.AspNetCore.App]
  Microsoft.AspNetCore.App 2.1.3 [C:\Program Files\dotnet\shared\Microsoft.AspNetCore.App]
  Microsoft.NETCore.App 2.0.9 [C:\Program Files\dotnet\shared\Microsoft.NETCore.App]
  Microsoft.NETCore.App 2.1.2 [C:\Program Files\dotnet\shared\Microsoft.NETCore.App]
  Microsoft.NETCore.App 2.1.3 [C:\Program Files\dotnet\shared\Microsoft.NETCore.App]