基于Dapper打造高性能PostgreSQL异步访问封装(DapperHelper PostgerSQL 版)
一、前言
在高富贵性的环境中,我们常常需要处理大量数据,而Dapper作为一款轻量级ORM,充分发挥了性能上的优势。
然而,默认的Dapper没有考虑大量数据流式处理,很容易一次性拉入大量数据到内存,造成OOM。
本文将教你如何封装一个完全支持「异步」「流式」「多结果集」的 DapperHelper,让你充分采用PostgreSQL和Dapper的性能优势!
二、Dapper和PostgreSQL简介
- Dapper:定位于极致性能的微型ORM,直接将SQL查询结果映射成实体。
- PostgreSQL:一款免费、开源、强大的具有高实时性性能的关系型数据库,特别适合处理大量数据库操作。
三、DapperHelper设计思路
- 异步打开连接 (OpenAsync)
- 全部方法完全异步
- 流式读取 (StreamQueryAsync)
- 多结果集支持 (QueryMultipleAsync)
四、流式、多结果封装
public static async IAsyncEnumerable<T> StreamQueryAsync<T>(string sql, object? parameters = null)
{
await using var connection = await CreateConnectionAsync();
await using var reader = await connection.ExecuteReaderAsync(sql, parameters, commandType: CommandType.Text, commandBehavior: CommandBehavior.SequentialAccess | CommandBehavior.CloseConnection);
var parser = reader.GetRowParser<T>();
while (await reader.ReadAsync())
{
yield return parser(reader);
}
}
public static async Task<TResult> QueryMultipleAsync<TResult>(string sql, object? parameters, Func<SqlMapper.GridReader, Task<TResult>> mapFunc)
{
await using var connection = await CreateConnectionAsync();
using var multi = await connection.QueryMultipleAsync(sql, parameters);
return await mapFunc(multi);
}
这样的设计,可以保证日常工作中:
- 处理大表查询不爆内存
- 多次SQL处理简单、开放
五、性能比较
| 方式 | 内存占用 | 速度 | 适用场景 |
|---|---|---|---|
| QueryAsync(默认) | 高 | 快 | 小量数据(<10万条) |
| StreamQueryAsync | 低 | 较快 | 大表,流水统计查询 |
六、DapperHelper完整代码
using System;
using System.Collections.Generic;
using System.Data;
using System.Threading.Tasks;
using Dapper;
using Npgsql;
namespace YourNamespace
{
/// <summary>
/// 提供对 PostgreSQL 数据库的 Dapper 异步访问封装,支持常规查询、事务操作、流式查询和多结果集查询。
/// </summary>
public static class DapperHelper
{
private static readonly string _connectionString = "Host=your_host;Port=5432;Database=your_database;Username=your_user;Password=your_password;Pooling=true;Maximum Pool Size=100;";
/// <summary>
/// 创建并打开一个异步 PostgreSQL 数据库连接。
/// </summary>
/// <returns>已打开的 NpgsqlConnection 实例</returns>
private static async Task<NpgsqlConnection> CreateConnectionAsync()
{
var connection = new NpgsqlConnection(_connectionString);
await connection.OpenAsync();
return connection;
}
/// <summary>
/// 执行异步 SQL 查询并返回结果集。
/// </summary>
/// <typeparam name="T">映射的实体类型</typeparam>
/// <param name="sql">SQL 查询语句</param>
/// <param name="parameters">参数对象</param>
/// <returns>实体集合</returns>
public static async Task<IEnumerable<T>> QueryAsync<T>(string sql, object? parameters = null)
{
await using var connection = await CreateConnectionAsync();
return await connection.QueryAsync<T>(sql, parameters);
}
/// <summary>
/// 执行异步 SQL 查询并返回第一条记录或默认值。
/// </summary>
/// <typeparam name="T">映射的实体类型</typeparam>
/// <param name="sql">SQL 查询语句</param>
/// <param name="parameters">参数对象</param>
/// <returns>单个实体或 null</returns>
public static async Task<T?> QueryFirstOrDefaultAsync<T>(string sql, object? parameters = null)
{
await using var connection = await CreateConnectionAsync();
return await connection.QueryFirstOrDefaultAsync<T>(sql, parameters);
}
/// <summary>
/// 执行 INSERT/UPDATE/DELETE 等命令。
/// </summary>
/// <param name="sql">SQL 命令</param>
/// <param name="parameters">参数对象</param>
/// <returns>受影响的行数</returns>
public static async Task<int> ExecuteAsync(string sql, object? parameters = null)
{
await using var connection = await CreateConnectionAsync();
return await connection.ExecuteAsync(sql, parameters);
}
/// <summary>
/// 执行带事务的操作。
/// </summary>
/// <param name="action">包含连接与事务的异步操作</param>
/// <returns>执行结果(返回值由调用者决定)</returns>
public static async Task<int> ExecuteTransactionAsync(Func<IDbConnection, IDbTransaction, Task<int>> action)
{
await using var connection = await CreateConnectionAsync();
await using var transaction = await connection.BeginTransactionAsync();
try
{
var result = await action(connection, transaction);
await transaction.CommitAsync();
return result;
}
catch
{
await transaction.RollbackAsync();
throw;
}
}
/// <summary>
/// 批量执行多条 SQL 命令,并在同一事务中提交。
/// </summary>
/// <param name="commands">SQL 命令与参数的集合</param>
/// <returns>是否成功提交事务</returns>
public static async Task<bool> ExecuteBatchAsync(IEnumerable<(string Sql, object? Parameters)> commands)
{
await using var connection = await CreateConnectionAsync();
await using var transaction = await connection.BeginTransactionAsync();
try
{
foreach (var (sql, parameters) in commands)
{
await connection.ExecuteAsync(sql, parameters, transaction);
}
await transaction.CommitAsync();
return true;
}
catch
{
await transaction.RollbackAsync();
throw;
}
}
/// <summary>
/// 异步流式读取数据,适用于超大数据集。
/// </summary>
/// <typeparam name="T">映射的实体类型</typeparam>
/// <param name="sql">SQL 查询语句</param>
/// <param name="parameters">参数对象</param>
/// <returns>可异步枚举的数据流</returns>
public static async IAsyncEnumerable<T> StreamQueryAsync<T>(string sql, object? parameters = null)
{
await using var connection = await CreateConnectionAsync();
await using var reader = await connection.ExecuteReaderAsync(sql, parameters, commandType: CommandType.Text, commandBehavior: CommandBehavior.SequentialAccess | CommandBehavior.CloseConnection);
var parser = reader.GetRowParser<T>();
while (await reader.ReadAsync())
{
yield return parser(reader);
}
}
/// <summary>
/// 执行 SQL 并读取多个结果集(多表查询)。
/// </summary>
/// <typeparam name="TResult">最终组合返回类型</typeparam>
/// <param name="sql">SQL 查询语句,可包含多个 SELECT</param>
/// <param name="parameters">参数对象</param>
/// <param name="mapFunc">结果集处理函数,接收 GridReader 并返回 TResult</param>
/// <returns>映射后的结果</returns>
public static async Task<TResult> QueryMultipleAsync<TResult>(
string sql,
object? parameters,
Func<SqlMapper.GridReader, Task<TResult>> mapFunc)
{
await using var connection = await CreateConnectionAsync();
using var multi = await connection.QueryMultipleAsync(sql, parameters);
return await mapFunc(multi);
}
}
}
通过自己封装一套 DapperHelper:
- 流式处理大表,避免内存爆表
- 多结果集处理灵活高效,容易扩展
在实际开发中,很推荐将这样的基础封装统一化,有效地拆分合理、降低Bug率,进而提升团队效率!
七、方法使用示例
QueryAsync 示例
var users = await DapperHelper.QueryAsync<User>(
"SELECT * FROM users WHERE age > @age",
new { age = 30 });
foreach (var user in users)
{
Console.WriteLine($"{user.Id} - {user.Name}");
}
QueryFirstOrDefaultAsync 示例
var user = await DapperHelper.QueryFirstOrDefaultAsync<User>(
"SELECT * FROM users WHERE id = @id",
new { id = 1 });
Console.WriteLine(user?.Name ?? "未找到用户");
ExecuteAsync 示例
var rowsAffected = await DapperHelper.ExecuteAsync(
"UPDATE users SET name = @name WHERE id = @id",
new { id = 1, name = "新名字" });
Console.WriteLine($"更新了 {rowsAffected} 条记录");
ExecuteTransactionAsync 示例
var count = await DapperHelper.ExecuteTransactionAsync(async (conn, tran) =>
{
var insertSql = "INSERT INTO logs(message) VALUES(@message)";
var deleteSql = "DELETE FROM users WHERE id = @id";
await conn.ExecuteAsync(insertSql, new { message = "删除用户前日志" }, tran);
return await conn.ExecuteAsync(deleteSql, new { id = 5 }, tran);
});
Console.WriteLine($"事务内总计受影响行数:{count}");
ExecuteBatchAsync 示例
var commands = new List<(string Sql, object? Parameters)>
{
("INSERT INTO products(name, price) VALUES(@name, @price)", new { name = "商品1", price = 99 }),
("INSERT INTO products(name, price) VALUES(@name, @price)", new { name = "商品2", price = 199 })
};
var success = await DapperHelper.ExecuteBatchAsync(commands);
Console.WriteLine(success ? "批量执行成功" : "批量执行失败");
StreamQueryAsync 示例(流式读取)
await foreach (var log in DapperHelper.StreamQueryAsync<LogEntry>(
"SELECT * FROM logs WHERE created_at > @time",
new { time = DateTime.UtcNow.AddDays(-1) }))
{
Console.WriteLine($"{log.Id}: {log.Message}");
}
QueryMultipleAsync 示例(多结果集)
var sql = @"
SELECT * FROM users WHERE id = @id;
SELECT * FROM orders WHERE user_id = @id;
";
var result = await DapperHelper.QueryMultipleAsync(sql, new { id = 1 }, async reader =>
{
var user = await reader.ReadFirstOrDefaultAsync<User>();
var orders = (await reader.ReadAsync<Order>()).ToList();
return (user, orders);
});
Console.WriteLine($"用户:{result.user?.Name}");
foreach (var order in result.orders)
{
Console.WriteLine($"订单号:{order.Id} 金额:{order.Amount}");
}
八、注意事项
✅ 必须安装的包
| 包名 |
|---|
| Dapper |
| Npgsql |
🚀 最低兼容环境要求
- .NET Core 3.1+ 或 .NET 5/6/7/8
- PostgreSQL 9.6+(一般建议 12及以上)
- Dapper 和 Npgsql 最新稳定版兼容良好
📢 小提醒
使用连接字符串,确保格式正确,例如:
Host=localhost;Port=5432;Database=mydb;Username=myuser;Password=mypassword;Pooling=true;Maximum Pool Size=100;
并且在生产环境中注意:
- 启用 连接池(Pooling=true)
- 设置合理的 最大连接数(Maximum Pool Size)
欢迎收藏,点赞,评论与分享!✨
如需要进一步扩展版(含Redis缓存/分表处理/分流分表处理等),可留言或私信我!
241

被折叠的 条评论
为什么被折叠?



