前言：这两天重温经典，对ado.net的东西稍微深入的了解了一下，顺便写点代码练练手，全当是复习笔记吧。
一、简单说说ado.net的5大常用对象
既然说ado.net，当然不能免俗地要提到5大常用对象。本文不会对ado.net的5大对象和它们的关系进行过多阐释，不过我们应该对下面这张图的结构有个了解：
关于上图图示中的5大对象，经常做以数据为驱动的mis系统的童鞋应该不会陌生。本文一笔带过。下面我们一步一步实现以ado.net为核心的数据访问程序。
【注意：下面的示例代码和demo是楼猪本周六和周日两天时间实现的，未经详细测试，可能有重大bug，下载学习使用的童鞋务必注意】
二、数据访问持久化层
1、IDbOperation接口
代码
using System.Collections.Generic;
using System.Data;
using System.Data.Common;
namespace AdoNetDataaccess.Core.Contract
{
public interface IDbOperation
{
DbCommand CreateDbCommd(DbConnection sqlConn, DbTransaction transaction, string sqlStr, CommandType cmdType, List<DbParameter> listParams);
DbParameter CreateDbPRameter(string paramName, object paramValue);
DbDataReader ExecuteReader(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
DataTable FillDataTable(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
DataSet FillDataSet(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
object ExecuteScalar(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
int ExecuteNonQuery(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
/// <summary>
/// 批量插入
/// </summary>
/// <param name="tableName">表名称</param>
/// <param name="dt">组装好的要批量导入的datatable</param>
/// <returns></returns>
bool ExecuteBatchInsert(string tableName, int batchSize, int copyTimeout, DataTable dt);
void OpenConnection();
void CloseConnection();
}
}
上面的接口包括增删改查，批量插入以及数据库连接对象的连接和关闭等常用操作，您可以根据命名和参数轻松理解函数的含义。根据楼猪的开发经验，对于平时的数据库操作，上述方法差不多够用了。当然您也可以按照自己需要，重写组织添加其他函数。
2、针对一种数据源的数据操作实现
底层的数据操作接口定义好后，就要针对一种数据源，具体实现上述的数据操作。这里楼猪选择了Sql Server。我们也可以实现其他数据源的数据访问操作，按照配置，利用抽象工厂动态反射选择是哪一种数据源的实现。这里按下不表，有心的童鞋自己可以动手一试。下面是具体的实现：
代码
using System;
using System.Collections.Generic;
using System.Data;
using System.Data.Common;
using System.Data.SqlClient;
using System.Transactions;
namespace AdoNetDataAccess.Core.Implement
{
using AdoNetDataAccess.Core.Contract;
public class SqlServer : IDbOperation, IDisposable
{
private int cmdTimeOut = 60;
private DbConnection sqlConn = null;
private DbCommand cmd = null;
private SqlServer()
{
}
public SqlServer(string sqlConStr)
{
sqlConn = new SqlConnection(sqlConStr);
cmdTimeOut = sqlConn.ConnectionTimeout;
}
public SqlServer(string sqlConStr, int timeOut)
{
sqlConn = new SqlConnection(sqlConStr);
if (timeOut < 0)
{
timeOut = sqlConn.ConnectionTimeout;
}
cmdTimeOut = timeOut;
}
#region contract method
public DbCommand CreateDbCommd(DbConnection sqlConn, DbTransaction transaction, string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
DbCommand cmd = new SqlCommand();
cmd.Connection = sqlConn;
cmd.CommandText = sqlStr;
cmd.CommandType = cmdType;
if (transaction != null)
{
cmd.Transaction = transaction;
}
if (listParams != null && listParams.Count > 0)
{
cmd.Parameters.AddRange(listParams.ToArray());
}
cmd.CommandTimeout = cmdTimeOut;
OpenConnection();
return cmd;
}
public DbParameter CreateDbPrameter(string paramName, object paramValue)
{
SqlParameter sp = new SqlParameter(paramName, paramValue);
return sp;
}
public DbDataReader ExecuteReader(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
DbDataReader rdr = null;
try
{
OpenConnection();
cmd = CreateDbCommd(sqlConn, null, sqlStr, cmdType, listParams);
rdr = cmd.ExecuteReader();
}
catch (Exception ex)
{
throw ex;
}
return rdr;
}
public DataTable FillDataTable(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
OpenConnection();
DbTransaction trans = sqlConn.BeginTransaction();
DbCommand cmd = CreateDbCommd(sqlConn, trans, sqlStr, cmdType, listParams);
SqlDataAdapter sqlDataAdpter = new SqlDataAdapter(cmd as SqlCommand);
DataTable dt = new DataTable();
try
{
sqlDataAdpter.Fill(dt);
trans.Commit();
}
catch (Exception e)
{
trans.Rollback();
throw new Exception("执行数据库操作失败, sql: " + sqlStr, e);
}
finally
{
sqlDataAdpter.Dispose();
cmd.Dispose();
trans.Dispose();
CloseConnection();
}
return dt;
}
public DataSet FillDataSet(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
OpenConnection();
DbTransaction trans = sqlConn.BeginTransaction();
DbCommand cmd = CreateDbCommd(sqlConn, trans, sqlStr, cmdType, listParams);
SqlDataAdapter sqlDataAdpter = new SqlDataAdapter(cmd as SqlCommand);
DataSet ds = new DataSet();
try
{
sqlDataAdpter.Fill(ds);
trans.Commit();
}
catch (Exception e)
{
trans.Rollback();
throw new Exception("执行数据库操作失败, sql: " + sqlStr, e);
}
finally
{
sqlDataAdpter.Dispose();
cmd.Dispose();
trans.Dispose();
CloseConnection();
}
return ds;
}
public object ExecuteScalar(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
object result = null;
OpenConnection();
DbTransaction trans = sqlConn.BeginTransaction();
try
{
cmd = CreateDbCommd(sqlConn, trans, sqlStr, cmdType, listParams);
result = cmd.ExecuteScalar();
trans.Commit();
}
catch (Exception e)
{
trans.Rollback();
throw new Exception("执行数据库操作失败, sql: " + sqlStr, e);
}
finally
{
trans.Dispose();
CloseConnection();
}
return result;
}
public int ExecuteNonQuery(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
int result = -1;
OpenConnection();
DbTransaction trans = sqlConn.BeginTransaction();
try
{
cmd = CreateDbCommd(sqlConn, trans, sqlStr, cmdType, listParams);
result = cmd.ExecuteNonQuery();
trans.Commit();
}
catch (Exception e)
{
trans.Rollback();
throw new Exception("执行数据库操作失败, sql: " + sqlStr, e);
}
finally
{
trans.Dispose();
CloseConnection();
}
return result;
}
/// <summary>
/// 批量插入
/// </summary>
/// <param name="tableName"></param>
/// <param name="batchSize"></param>
/// <param name="copyTimeout"></param>
/// <param name="dt"></param>
/// <returns></returns>
public bool ExecuteBatchInsert(string tableName, int batchSize, int copyTimeout, DataTable dt)
{
bool flag = false;
try
{
using (TransactionScope scope = new TransactionScope())
{
OpenConnection();
using (SqlBulkCopy sbc = new SqlBulkCopy(sqlConn as SqlConnection))
{
//服务器上目标表的名称
sbc.DestinationTableName = tableName;
sbc.BatchSize = batchSize;
sbc.BulkCopyTimeout = copyTimeout;
for (int i = 0; i < dt.Columns.Count; i++)
{
//列映射定义数据源中的列和目标表中的列之间的关系
sbc.ColumnMappings.Add(dt.Columns[i].ColumnName, dt.Columns[i].ColumnName);
}
sbc.WriteToServer(dt);
flag = true;
scope.Complete();//有效的事务
}
}
}
catch (Exception ex)
{
throw ex;
}
return flag;
}
public void OpenConnection()
{
if (sqlConn.State == ConnectionState.Broken || sqlConn.State == ConnectionState.Closed)
sqlConn.Open();
}
public void CloseConnection()
{
sqlConn.Close();
}
#endregion
#region dispose method
/// <summary>
/// dispose接口方法
/// </summary>
public void Dispose()
{
}
#endregion
}
}
到这里，我们实现了SqlServer类里的方法，对Ms SqlServer数据库我们就已经可以进行简单的基础的CRUD操作了。
三、简单直观的对象实体转换
在第二步中，我们已经实现了简单的数据CRUD操作。根据楼猪使用ORM的经验和习惯，我们也应该对一些查询结果进行转换，因为以类的组织方式比直接呈现ado.net对象更容易让人接受，效率高低反在其次。下面利用常见的反射原理，简单实现一个对象实体转换器ModelConverter类：
代码
using System;
using System.Collections;
using System.Collections.Generic;
using System.Data;
using System.Data.Common;
using System.Reflection;
using System.Threading;
namespace AdoNetDataAccess.Core.Obj2Model
{
using AdoNetDataAccess.Core.Contract;
public sealed class ModelConverter
{
private static readonly object objSync = new object();
#region query for list
/// <summary>
/// 查询数据表项并转换为对应实体
/// </summary>
/// <typeparam name="T"></typeparam>
/// <param name="objType"></param>
/// <param name="rdr"></param>
/// <returns></returns>
public static IList<T> QueryForList<T>(string sqlStr, CommandType cmdType, List<DbParameter> listParams, Type objType, IDbOperation dbOperation)
where T : class, new()
{
IDataReader rdr = dbOperation.ExecuteReader(sqlStr, cmdType, listParams);
IList<T> listModels = new List<T>();
try
{
Monitor.Enter(objSync);
Hashtable ht = CreateHashColumnName(rdr);
while (rdr.Read())
{
Object obj = Activator.CreateInstance(objType);
PropertyInfo[] properties = objType.GetProperties();
foreach (PropertyInfo propInfo in properties)
{
string columnName = propInfo.Name.ToUpper();
if (ht.ContainsKey(columnName) == false)
{
continue;
}
int index = rdr.GetOrdinal(propInfo.Name);
object columnValue = rdr.GetValue(index);
if (columnValue != System.DBNull.Value)
{
SetValue(propInfo, obj, columnValue);
}
}
T model = default(T);
model = obj as T;
listModels.Add(model);
}
}
finally
{
rdr.Close();
rdr.Dispose();
Monitor.Exit(objSync);
}
return listModels;
}
#endregion
#region query for dictionary
/// <summary>
/// 查询数据表项并转换为对应实体
/// </summary>
/// <typeparam name="K"></typeparam>
/// <typeparam name="T"></typeparam>
/// <param name="key">字典对应key列名</param>
/// <param name="objType"></param>
/// <param name="rdr"></param>
/// <returns></returns>
public static IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, CommandType cmdType, List<DbParameter> listParams, Type objType, IDbOperation dbOperation)
where T : class, new()
{
IDataReader rdr = dbOperation.ExecuteReader(sqlStr, cmdType, listParams);
IDictionary<K, T> dictModels = new Dictionary<K, T>();
try
{
Monitor.Enter(objSync);
Hashtable ht = CreateHashColumnName(rdr);
while (rdr.Read())
{
Object obj = Activator.CreateInstance(objType);
PropertyInfo[] properties = objType.GetProperties();
object dictKey = null;
foreach (PropertyInfo propInfo in properties)
{
string columnName = propInfo.Name.ToUpper();
if (ht.ContainsKey(columnName) == false)
{
continue;
}
int index = rdr.GetOrdinal(propInfo.Name);
object columnValue = rdr.GetValue(index);
if (columnValue != System.DBNull.Value)
{
SetValue(propInfo, obj, columnValue);
if (string.Compare(columnName, key.ToUpper()) == 0)
{
dictKey = columnValue;
}
}
}
T model = default(T);
model = obj as T;
K objKey = (K)dictKey;
dictModels.Add(objKey, model);
}
}
finally
{
rdr.Close();
rdr.Dispose();
Monitor.Exit(objSync);
}
return dictModels;
}
#endregion
#region internal util
private static Hashtable CreateHashColumnName(IDataReader rdr)
{
int len = rdr.FieldCount;
Hashtable ht = new Hashtable(len);
for (int i = 0; i < len; i++)
{
string columnName = rdr.GetName(i).ToUpper(); //不区分大小写
string columnRealName = rdr.GetName(i);
if (ht.ContainsKey(columnName) == false)
{
ht.Add(columnName, columnRealName);
}
}
return ht;
}
private static void SetValue(PropertyInfo propInfo, Object obj, object objValue)
{
try
{
propInfo.SetValue(obj, objValue, null);
}
catch
{
object realValue = null;
try
{
realValue = Convert.ChangeType(objValue, propInfo.PropertyType);
propInfo.SetValue(obj, realValue, null);
}
catch (Exception ex)
{
string err = ex.Message;
//throw ex; //在数据库数据有不符合规范的情况下应该及时抛出异常
}
}
}
#endregion
}
}
到这里，简单的数据访问持久化层就实现了。下面模仿楼猪使用的IBatis.net，写个伪SqlMapper，改善一下调用形式，丰富一下调用方法，让方法辨识度更高。
四、实现伪SqlMapper
1、BaseMapper类
代码
using System;
using System.Collections.Generic;
using System.Data;
using System.Data.Common;
using System.Data.SqlClient;
namespace AdoNetDataAccess.Mapper
{
using AdoNetDataAccess.Core.Contract;
public abstract class BaseMapper
{
public IDbOperation CurrentDbOperation;
#region query for list
public abstract IList<T> QueryForList<T>(string sqlStr)
where T : class, new();
public abstract IList<T> QueryForList<T>(string sqlStr, Type objType)
where T : class, new();
public abstract IList<T> QueryForList<T>(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
where T : class, new();
public abstract IList<T> QueryForList<T>(string sqlStr, CommandType cmdType, List<DbParameter> listParams, Type objType)
where T : class, new();
#endregion
#region query for dictionary
public abstract IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr)
where T : class, new();
public abstract IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, Type objType)
where T : class, new();
public abstract IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, CommandType cmdType, Type objType)
where T : class, new();
public abstract IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, CommandType cmdType, List<DbParameter> listParams, Type objType)
where T : class, new();
#endregion
#region dataset datatable
public abstract DataTable FillDataTable(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
public abstract DataSet FillDataSet(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
#endregion
#region ExecuteScalar
public abstract object ExecuteScalar(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
#endregion
#region insert
public abstract int Insert(string sqlStr);
public abstract int Insert(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
public abstract bool BatchInsert(string tableName, int batchSize, int copyTimeout, DataTable dt);
#endregion
#region delete
public abstract int Delete(string sqlStr);
public abstract int Delete(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
#endregion
#region update
public abstract int Update(string sqlStr);
public abstract int Update(string sqlStr, CommandType cmdType, List<DbParameter> listParams);
#endregion
}
}
上面代码中的方法您是不是很熟悉呢？ 呵呵，使用IBatis.net 的童鞋应该会和楼猪产生更多的共鸣。
2、SqlMapper类
代码
using System;
using System.Collections.Generic;
using System.Data;
using System.Data.Common;
namespace AdoNetDataAccess.Mapper
{
using AdoNetDataAccess.Core.Contract;
using AdoNetDataAccess.Core.Obj2Model;
public class SqlMapper : BaseMapper
{
private SqlMapper()
{
}
public SqlMapper(IDbOperation dbOperation)
{
this.CurrentDbOperation = dbOperation;
}
#region query for list
public override IList<T> QueryForList<T>(string sqlStr)
{
return QueryForList<T>(sqlStr, CommandType.Text, null, typeof(T));
}
public override IList<T> QueryForList<T>(string sqlStr, Type objType)
{
return QueryForList<T>(sqlStr, CommandType.Text, null, objType);
}
public override IList<T> QueryForList<T>(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return QueryForList<T>(sqlStr, cmdType, listParams, typeof(T));
}
public override IList<T> QueryForList<T>(string sqlStr, CommandType cmdType, List<DbParameter> listParams, Type objType)
{
return ModelConverter.QueryForList<T>(sqlStr, cmdType, listParams, objType, this.CurrentDbOperation);
}
#endregion
#region query for dictionary
public override IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr)
{
return QueryForDictionary<K, T>(key, sqlStr, CommandType.Text, null, typeof(T));
}
public override IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, Type objType)
{
return QueryForDictionary<K, T>(key, sqlStr, CommandType.Text, null, objType);
}
public override IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, CommandType cmdType, Type objType)
{
return QueryForDictionary<K, T>(key, sqlStr, cmdType, null, objType);
}
public override IDictionary<K, T> QueryForDictionary<K, T>(string key, string sqlStr, CommandType cmdType, List<DbParameter> listParams, Type objType)
{
return ModelConverter.QueryForDictionary<K, T>(key, sqlStr, cmdType, listParams, objType, this.CurrentDbOperation);
}
#endregion
#region dataset datatable
public override DataTable FillDataTable(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return this.CurrentDbOperation.FillDataTable(sqlStr, cmdType, listParams);
}
public override DataSet FillDataSet(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return this.CurrentDbOperation.FillDataSet(sqlStr, cmdType, listParams);
}
#endregion
#region ExecuteScalar
public override object ExecuteScalar(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return this.CurrentDbOperation.ExecuteScalar(sqlStr, cmdType, listParams);
}
#endregion
#region insert
public override int Insert(string sqlStr)
{
object obj = ExecuteScalar(sqlStr, CommandType.Text, null);
int id = obj == null ? 0 : int.Parse(obj.ToString());
return id;
}
public override int Insert(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
object obj = ExecuteScalar(sqlStr, cmdType, listParams);
int id = obj == null ? 0 : int.Parse(obj.ToString());
return id;
}
/// <summary>
/// 批量插入
/// </summary>
/// <param name="tableName"></param>
/// <param name="batchSize"></param>
/// <param name="copyTimeout"></param>
/// <param name="dt"></param>
/// <returns></returns>
public override bool BatchInsert(string tableName, int batchSize, int copyTimeout, DataTable dt)
{
return this.CurrentDbOperation.ExecuteBatchInsert(tableName, batchSize, copyTimeout, dt);
}
#endregion
#region delete
public override int Delete(string sqlStr)
{
return CommitSql(sqlStr, CommandType.Text, null);
}
public override int Delete(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return CommitSql(sqlStr, cmdType, listParams);
}
#endregion
#region update
public override int Update(string sqlStr)
{
return CommitSql(sqlStr, CommandType.Text, null);
}
public override int Update(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return CommitSql(sqlStr, cmdType, listParams);
}
#endregion
#region commit and execute sql
private int CommitSql(string sqlStr, CommandType cmdType, List<DbParameter> listParams)
{
return this.CurrentDbOperation.ExecuteNonQuery(sqlStr, cmdType, listParams);
}
#endregion
#region dbparameter
public DbParameter CreateParameter(string paraName, object paramValue)
{
return this.CurrentDbOperation.CreateDbPrameter(paraName, paramValue);
}
public List<DbParameter> CreateParameterList(string[] paraNames, object[] paramValues)
{
List<DbParameter> listParams = new List<DbParameter>();
try
{
if (paraNames.Length != paramValues.Length)
{
throw new Exception("Param name and value is not equal.");
}
for (int i = 0; i < paraNames.Length; i++)
{
DbParameter param = CreateParameter(paraNames[i], paramValues[i]);
listParams.Add(param);
}
}
catch (Exception ex)
{
throw ex;
}
return listParams;
}
#endregion
}
}
上面的方法丰富实现了CRUD的常见操作，其实主要还是调用了IDbOperation接口和方法。
未完，待续。
demo下载：demo
-
