function that can use iText to concatenate / merge pdfs together - causing some issues

Question

I'm using the following code to merge PDFs together using iText:

public static void concatenatePdfs(List<File> listOfPdfFiles, File outputFile) throws DocumentException, IOException {
        Document document = new Document();
        FileOutputStream outputStream = new FileOutputStream(outputFile);
        PdfWriter writer = PdfWriter.getInstance(document, outputStream);
        document.open();
        PdfContentByte cb = writer.getDirectContent();
        for (File inFile : listOfPdfFiles) {
            PdfReader reader = new PdfReader(inFile.getAbsolutePath());
            for (int i = 1; i <= reader.getNumberOfPages(); i++) {
                document.newPage();
                PdfImportedPage page = writer.getImportedPage(reader, i);
                cb.addTemplate(page, 0, 0);
            }
        }
        outputStream.flush();
        document.close();
        outputStream.close();
    }

This usually works great! But once and a while, it's rotating some of the pages by 90 degrees? Anyone ever have this happen?

I am looking into the PDFs themselves to see what is special about the ones that are being flipped.

score 18 · Accepted Answer · answered Apr 14 '14 at 15:04

18

There are errors once in a while because you are using the wrong method to concatenate documents. Please read chapter 6 of my book and you'll notice that using PdfWriter to concatenate (or merge) PDF documents is wrong:

You completely ignore the page size of the pages in the original document (you assume they are all of size A4),
You ignore page boundaries such as the crop box (if present),
You ignore the rotation value stored in the page dictionary,
You throw away all interactivity that is present in the original document, and so on.

Concatenating PDFs is done using PdfCopy, see for instance the FillFlattenMerge2 example:

Document document = new Document();
PdfCopy copy = new PdfSmartCopy(document, new FileOutputStream(dest));
document.open();
PdfReader reader;
String line = br.readLine();
// loop over readers
    // add the PDF to PdfCopy
    reader = new PdfReader(baos.toByteArray());
    copy.addDocument(reader);
    reader.close();
// end loop
document.close();

There are other examples in the book.

answered Apr 14 '14 at 15:04

Bruno Lowagie

75,994
9
109
165

2

I recognize your name from the itext namespaces. Are you a founder of the itext library? – Nicholas DiPiazza Apr 15 '14 at 18:18
8

Yes, I'm that Lowagie ;-) – Bruno Lowagie Apr 15 '14 at 18:21
Question - copy.addDocument doesn't exist in my version of itext. am I too old? 5.0.6 – Nicholas DiPiazza Apr 15 '14 at 18:23
1

yep 5.5.0 worked. i thought i was grabbing latest from maven repository but i accidentally mis clicked when i went to copy paste. crap. – Nicholas DiPiazza Apr 15 '14 at 18:26
Now a better question - why didn't you have to close your file output stream? Does document.close() do that for you? – Nicholas DiPiazza Apr 15 '14 at 18:27
1

Yes, `document.close()` also closes the `PdfWriter` instance and the `OutputStream` used by the writer. You can prevent this by using `writer.setCloseStream(false);`. For instance: if you have a `ZipOutputStream` to which you're adding multiple PDFs, you don't want iText to close that stream. – Bruno Lowagie Apr 16 '14 at 06:56
The code where I'm calling .close() seems to work, but you're saying then I don't need it. OK. Thanks for the great info. I have confirmed the rotating pages are gone now with your help. – Nicholas DiPiazza Apr 16 '14 at 18:02
@NicholasDiPiazza, if you have found your own solution, even using another answer, make your own answer, don't try to [modify another answer](http://stackoverflow.com/review/suggested-edits/4592838). – gunr2171 Apr 16 '14 at 18:04
more of an attempt to add on to his answer the correct version of the function i attempted to write above. I can make as a new answer – Nicholas DiPiazza Apr 16 '14 at 18:04

Nicholas DiPiazza · Answer 2 · 2014-04-16T18:10:40.917

In case anyone is looking for it, using Bruno Lowagie's correct answer above, here is the version of the function that does not seem to have the page flipping issue i described above:

public static void concatenatePdfs(List<File> listOfPdfFiles, File outputFile) throws DocumentException, IOException {
        Document document = new Document();
        FileOutputStream outputStream = new FileOutputStream(outputFile);
        PdfCopy copy = new PdfSmartCopy(document, outputStream);
        document.open();
        for (File inFile : listOfPdfFiles) {
            PdfReader reader = new PdfReader(inFile.getAbsolutePath());
            copy.addDocument(reader);
            reader.close();
        }
        document.close();
}

function that can use iText to concatenate / merge pdfs together - causing some issues

2 Answers2

Linked