Creating PDF with overflow layout via iTextSharp

Question

I'm successfully creating a PDF document from a DataTable via iTextSharp, but I can't get the layout in a desireable format.

Though unlikely, the DataTable has the potential for dozens of columns. Imagine the following DataTable - each number represents the area that can fit on a page:

|------------|
| 1 : 2 : 3  |
|------------|
| 4 : 5 : 6  |
|------------|

This should export as a 6 page PDF document in the order I've numbered the sections. Instead, it's currently generating as a two page document with 40-some columns squished on each page with a font so small it literally has no detail.

The simplest way to explain what I want is: When a generated PDF prints, I want it to print like a very wide Excel sheet rather than squishing all the content together.

My code is as follows:

    public static void ExportAsPDF(DataTable Table, IList<string> Columns, string filename)
    {
        int ColumnCount = Table.Columns.Count;
        int RowCount = Table.Rows.Count;

        iTextSharp.text.Table BodyTable = new iTextSharp.text.Table(ColumnCount, RowCount);
        BodyTable.AutoFillEmptyCells = true;

        foreach (string s in Columns)
        {
            BodyTable.AddCell(s);
        }
        foreach (object o in from DataRow row in Table.Rows from o in row.ItemArray select o)
        {
            BodyTable.AddCell(o.ToString());
        }

        Document doc = new Document();
        PdfWriter.GetInstance(doc, HttpContext.Current.Response.OutputStream);
        doc.Open();
        doc.Add(BodyTable);
        doc.Close();

        HttpContext.Current.Response.ContentType = "application/pdf";
        HttpContext.Current.Response.AddHeader("content-disposition", string.Format("attachment; filename={0}.pdf", filename));
        HttpContext.Current.Response.End(); 
    }

All input and help is much appreciated. Thanks

Answer 1

Okay, two options.

1. Draw everything yourself by hand with PdfContentByte and ColumnText for the text layout part.

2. Convince the normal iText Layout code that your page is wide enough to hold an entire row, then split those pages into pieces later. Not Pretty, but probably easier than the alternative.

First you need to define a page that is 3 times wider than normal.

Rectangle triplePageRect = new Rectangle(PageSize.LETTER);
float origWidth = triplePageRect.getWidth();
triplePageRect.setWidth(origWidth * 3f);

Document doc = new Document(triplePageRect);
ByteArrayOutputStream baos = new ByteArrayOutputStream();
PdfWriter writer = PdfWriter.getInstance(doc, baos);

Then you draw your table pretty much the way you are now... BUT you must be sure that a column edge lines up with your two page edges so you can easily break up the pages later. Bonus points if you can create an empty row centered on where the page break will be so you have a margin.

//Your Code Here

Finally, you need to save your PDF, open it again, and chop it to ribbons. I'm thinking PdfStamper.

// write everything out to the baos.
doc.close(); 

// and suck it right back up again.  Hurray for efficiency.  Or something.
PdfReader reader = new PdfReader(baos.toByteArrayOrWhateverItsCalled());
PdfStamper stamper = new PdfStamper( reader, new FileOutputStream(outputPath));

// duplicate the pages.  I'll assume only one page, but you'll get the idea.
PdfDictionary origPage = reader.getPageN(1);
for (int i = 0; i < 2; ++i) {
  // initial size is irrelevant, we're going to change it, but !null
  stamper.insertPage(2+i, PageSize.LETTER); 
  PdfDictionary newPageDict = reader.getPage(2 + i);

  // copy the original page... note that this is a shallow copy
  newPageDict.putAll(origPageDict);

  // duplicate the page rect so each page will have its own copy
  PdfArray pageRect = newPageDict.getAsArray(PdfName.MEDIABOX);
  // also a shallow copy, but changes to this array will be localized to the page.
  PdfArray newRect = new PdfArray(pageRect);
  // page rects are defined as [llx lly urx ury], so we need to change 0 and 2.
  newRect.set(0, new PdfNumber(origWidth * (i+1));
  newRect.set(2, new PdfNumber(origWidth * (i+2));
}

//Smoke'em if you've got 'em folks, we're done.
stamper.close();

Damn I'm good. Oooh ooh! And modest! Let's not forget good looking, brave, courteous, witty...