QueryMultiple 结果集顺序已更改

Question

我正在使用 QueryMultiple 到 return 多组数据执行存储过程。

var gridReader = db.QueryMultiple("sp", 
                                  parameters, 
                                  commandType: CommandType.StoredProcedure);

如果我知道它们返回的顺序，我可以很容易地得到每套。

SELECT * FROM dbo.Set1;
SELECT * FROM dbo.Set2;
SELECT * FROM dbo.Set3;

var set1 = gridReader.Read<Set1>();
var set2 = gridReader.Read<Set2>();
var set3 = gridReader.Read<Set3>();

但是，我的情况是他们回来的顺序可能会改变。另一个开发人员可以出于任何原因决定更改顺序。存储过程现在变成这样：

SELECT * FROM dbo.Set1;
SELECT * FROM dbo.Set3;
SELECT * FROM dbo.Set2;

我该如何处理？

我最初的尝试是遍历每个网格，检查列名。起初这似乎运作良好，但除了手动设置每个字段之外，我无法弄清楚如何将网格投影到 class。我使用 Dapper 的主要原因是它可以为我做这件事。

while (true)
{
    var grid = gridReader.Read();
    IDictionary<string, object> row = grid.FirstOrDefault();

    if (row == null)
        break;

    if (row.Keys.Contains("Set1_UniqueColumnName"))
    {
        // Need something like grid.Read<Set1>();
    }
    else if (row.Keys.Contains("Set2_UniqueColumnName")) { }
    else if (row.Keys.Contains("Set3_UniqueColumnName")) { }
}

我的第二个想法是将每个网格读入 class，检查 class 的唯一字段中的 nulls/default 值，然后尝试下一个 class 如果测试失败。但这显然行不通。 .Read() 将 return 下一个 结果网格。此解决方案要求我能够一遍又一遍地读取相同的网格。

Answer 1

Dapper 提供了一个 IDataReader.GetRowParser 扩展方法，可以启用每行类型切换。来自 Dapper 文档 here...

Usually you'll want to treat all rows from a given table as the same data type. However, there are some circumstances where it's useful to be able to parse different rows as different data types. This is where IDataReader.GetRowParser comes in handy.

Imagine you have a database table named "Shapes" with the columns: Id, Type, and Data, and you want to parse its rows into Circle, Square, or Triangle objects based on the value of the Type column.

var shapes = new List<IShape>();
using (var reader = connection.ExecuteReader("select * from Shapes"))
{
    // Generate a row parser for each type you expect.
    // The generic type <IShape> is what the parser will return.
    // The argument (typeof(*)) is the concrete type to parse.
    var circleParser = reader.GetRowParser<IShape>(typeof(Circle));
    var squareParser = reader.GetRowParser<IShape>(typeof(Square));
    var triangleParser = reader.GetRowParser<IShape>(typeof(Triangle));

    var typeColumnIndex = reader.GetOrdinal("Type");

    while (reader.Read())
    {
        IShape shape;
        var type = (ShapeType)reader.GetInt32(typeColumnIndex);
        switch (type)
        {
            case ShapeType.Circle:
                shape = circleParser(reader);
                break;
            case ShapeType.Square:
                shape = squareParser(reader);
                break;
            case ShapeType.Triangle:
                shape = triangleParser(reader);
                break;
            default:
                throw new NotImplementedException();
        }

        shapes.Add(shape);
    }
}

您需要访问 GridReader 包装的 IDataReader 或更改您的代码以使用良好的老式 ADO.NET SqlConnection & SqlCommand 这样的对象...

using (command = new SqlCommand("sp", connection))
{
    command.CommandType = CommandType.StoredProcedure;
    command.Parameters.AddRange(parameters);

    using (var reader = command.ExecuteReader())
    {
        while (reader.Read())
        {
            // read row columns
        }
    }
}

Answer 2

给我指出了正确的方向。需要结合使用 ADO.NET 和 Dapper。本质上使用 ADO.NET 来检索和遍历数据，但使用 Dapper 将行解析为我的对象。请注意在 while 循环中使用 FieldCount 以防结果集实际执行 return 0 行。我们希望它移动到下一个结果集，而不是跳出循环。

Set1 set1 = null;
var set2 = new List<Set2>();
Set3 set3 = null;

using (var command = new SqlCommand("sp", conn))
{
    command.CommandType = CommandType.StoredProcedure;
    command.Parameters.AddRange(parameters);
    command.Connection.Open();

    using (var reader = command.ExecuteReader())
    {
        while (reader.FieldCount > 0)
        {
            var set1Parser = reader.GetRowParser<Set1>();
            var set2Parser = reader.GetRowParser<Set2>();
            var set3Parser = reader.GetRowParser<Set3>();

            var isSet1 = HasColumn(reader, "Set1_UniqueColumnName");
            var isSet2 = HasColumn(reader, "Set2_UniqueColumnName");
            var isSet3 = HasColumn(reader, "Set3_UniqueColumnName");

            while (reader.Read())
            {
                if (isSet1)
                {
                    set1 = set1Parser(reader);
                }
                else if (isSet2)
                {
                    set2.Add(set2Parser(reader));
                }
                else if (isSet3)
                {
                    set3 = set3Parser(reader);
                }
            }

            reader.NextResult();
        }
    }
}

public static bool HasColumn(IDataReader reader, string columnName)
{
    for (var i = 0; i < reader.FieldCount; i++)
    {
        if (reader.GetName(i).Equals(columnName, StringComparison.InvariantCultureIgnoreCase))
        {
            return true;
        }
    }

    return false;
}

QueryMultiple 结果集顺序已更改

QueryMultiple Result Set Order Changed

c#

dapper