Use iTextSharp to save a PDF to a SQL Server 2008 Blob, and read that Blob to save to disk

The need to deal with a pdf in multiple passes was not immediately clear when I first started working them, so maybe this is some help to you.

In the method below, we create a pdf, render it to a byte[], load it for post processing, render the pdf again and return the result.

The rest of your question deals with getting a byte[] into and out of a varbinary[max], saving a byte[] to file and reading it back out, which you can google easily enough.

public byte[] PdfGeneratorAndPostProcessor()
{
  byte[] newPdf;

  using (var pdf = new MemoryStream())
  using (var doc = new Document(iTextSharp.text.PageSize.A4))
  using (PdfWriter.GetInstance(doc, pdf))
  {
    doc.Open();

    // do stuff to the newly created doc...

    doc.Close();
    newPdf = pdf.GetBuffer();
  }      

  byte[] postProcessedPdf;
  var reader = new PdfReader(newPdf);

  using (var pdf = new MemoryStream())
  using (var stamper = new PdfStamper(reader, pdf))
  {
    var pageCount = reader.NumberOfPages;
    for (var i = 1; i <= pageCount; i++)
    {
      // do something on each page of the existing pdf
    }

    stamper.Close();
    postProcessedPdf = pdf.GetBuffer();
  }

  reader.Close();
  return postProcessedPdf;
}