C#

int startIndex = 0;
int buffersize = 2097152; //(1024*2048=2097152)
byte[] outbyte = new byte[buffersize];

long retval = r.GetBytes(2, startIndex, outbyte, 0, buffersize);
while (retval == buffersize)
{
  bw.Write(outbyte);
  if(startIndex%1048576==0) //FLush every MB? (1024*1024=1048576)
     bw.Flush(); 
  startIndex += buffersize;
  retval = r.GetBytes(WriterTabell.Reader.GetOrdinal("nef_elfil"), startIndex, outbyte, 0, buffersize);
}
bw.Write(outbyte, 0, (int)retval);

//flush, close, dispose

for(int iter=0; retval==buffersize; iter++) {
   ...
   if ((iter%16)==0) bw.Flush();
   ...
}

Thomas ST wrote:

any way I can find the optimal buffersize?

Thomas ST wrote:

Can .NET do the job for me (without using DataSet)?

using 
( 
    System.IO.FileStream fs 
= 
    fi.Open 
    (
        System.IO.FileMode.Create
    , 
        System.IO.FileAccess.Write
    ,   
        System.IO.FileShare.None
    )
)
{
    byte[] content = (byte[]) ds [ 1 ].Rows [ i ]  [ "FileContent" ] ;
                                    
    fs.Write ( content , 0 , content.Length ) ; 
                                
    fs.Close() ;
}

i call first an other running program

GWBas1c wrote:

Like I said, I'm deploying on Mono + Linux, so I can't use COM or CLR-specific APIs.