SQL Server的批量上传策略

Question

I am uploading data from CSV to SQL table using the following function.

Is there a better way to do it?

I am concerned about , right now, connection hold for long time. Hence need to be reduced.

   public bool SaveProxyBulkUploadData(List<AddServerPError> saveBulkUploadData)
        {
            try
            {
                foreach (AddServerPError addServetData in saveBulkUploadData)
                {
                    DbCommand dbCmd = CitiScriptExecutionDB.GetStoredProcCommand("USP_HS_InsertProxyBulkUploadData");
                    CitiScriptExecutionDB.AddInParameter(dbCmd, "@groupid", DbType.String, addServetData.GroupId);
                    CitiScriptExecutionDB.AddInParameter(dbCmd, "@proxyname", DbType.String, addServetData.ProxyName);
                    CitiScriptExecutionDB.AddInParameter(dbCmd, "@proxytype", DbType.String, addServetData.ProxyType);
                    CitiScriptExecutionDB.AddInParameter(dbCmd, "@sa_spoc_dl", DbType.String, addServetData.SA_SPOC_DL);
                    CitiScriptExecutionDB.AddInParameter(dbCmd, "@assignmentgroup", DbType.String, addServetData.AssignmentGroup);
                    CitiScriptExecutionDB.AddInParameter(dbCmd, "@createdby", DbType.String, addServetData.CreatedBy);
                    CitiScriptExecutionDB.ExecuteNonQuery(dbCmd);
                }

                return true;
            }
            catch (Exception ex)
            {
                string msg = ex.Message + ex.StackTrace;
                throw;
            }
        }

Stored Procedure

CREATE PROCEDURE USP_HS_InsertBulkUploadData   
 @groupid  int,  
 @hostname varchar(50),   
 @type1 varchar(50),  
 @type2 varchar(50),  
 @createdby varchar(50)       
AS        
BEGIN       

 Insert into [dbo].[EUCUsecaseGroupInputParam]    
 (    
  GroupId,  
  HostName,    
  Type1,    
  Type2,     
  CreatedBy,     
  CreatedDate,     
  UpdatedBy,    
  UpdatedDate,    
  IsActive    
  )        
 Values(   
  @groupId,  
  @hostname,    
  @type1,    
  @type2,    
  @createdby,    
  GetDate(),    
  null,    
  null,    
  1    
  )       
END 

Answer 1

Your best bet here is probably SqlBulkCopy , which throws raw TDS at the server very efficiently. SqlBulkCopy takes two types of input:

• DataTable
• IDataReader

So at that point you have 3 options:

• transform your List<AddServerPError> to a DataTable manually
• use something like FastMember to obtain an IDataReader over your existing list (an example of this is at the bottom of the project page: https://github.com/mgravell/fast-member ) - on nuget: https://www.nuget.org/packages/FastMember/
• ditch your existing list, and read the CSV directly as an IDataReader - there is a "LumenWorks" CsvReader that has always worked well for that; the closest nuget link I can see is https://www.nuget.org/packages/LumenWorksCsvReader/ (although this isn't the original, it has been extended)

Answer 2

You can consider using Cinchoo ETL , an open source library to do the bulk upload CSV file to database.

Option 1:

Load the CSV file straight to database

string connectionstring = @"#YOUR DB ConnectionString#";
using (SqlBulkCopy bcp = new SqlBulkCopy(connectionstring))
{
    using (var p = new ChoCSVReader("#YOUR CSV FILE#"))
    {
        bcp.DestinationTableName = "#TABLENAME#";
        bcp.EnableStreaming = true;
        bcp.BatchSize = 10000;
        bcp.BulkCopyTimeout = 0;
        bcp.NotifyAfter = 100;
        bcp.SqlRowsCopied += delegate (object sender, SqlRowsCopiedEventArgs e)
        {
            Console.WriteLine(e.RowsCopied.ToString("#,##0") + " rows copied.");
        };
        bcp.WriteToServer(p.AsDataReader());
    }
}

Option 2:

If the load of CSV done already and outputted as List<AddServerPError> , you still can upload them to database as below

List<AddServerPError> objs = # Your input objects #;

string connectionstring = @"#YOUR DB ConnectionString#";
using (SqlBulkCopy bcp = new SqlBulkCopy(connectionstring))
{
    bcp.DestinationTableName = "#TABLENAME#";
    bcp.EnableStreaming = true;
    bcp.BatchSize = 10000;
    bcp.BulkCopyTimeout = 0;
    bcp.NotifyAfter = 100;
    bcp.SqlRowsCopied += delegate (object sender, SqlRowsCopiedEventArgs e)
    {
        Console.WriteLine(e.RowsCopied.ToString("#,##0") + " rows copied.");
    };
    bcp.WriteToServer(objs.AsDataReader());
}