iText7 Error while reading cross reference table. table will be rebuilt. file position {0} cross reference entry in this xref subsection

Question

I have a simple function that is merging together pdfs from an ArrayList using iText7

            File pdfMerged = new File("merged.pdf");
            PdfDocument pdf = new PdfDocument(new PdfWriter(pdfMerged));
            PdfMerger merger = new PdfMerger(pdf);


            for (InputStream in : files) {
            //Add pages from the each document
                PdfDocument sourcePdf = new PdfDocument(new PdfReader(in));
                merger.merge(sourcePdf, 1, sourcePdf.getNumberOfPages());
                
                sourcePdf.close();
            }

            pdf.close();
}

While this works most of the time, unfortunately one pdf is giving the following error.

2020-08-28 18:05:59,935  ERROR [kernel.pdf.PdfReader] [http-nio-8080-exec-9] Error occurred while reading cross reference table. Cross reference table will be rebuilt.
com.itextpdf.io.IOException: Error at file pointer 1,051,972.
at com.itextpdf.io.source.PdfTokenizer.throwError(PdfTokenizer.java:639)
at com.itextpdf.kernel.pdf.PdfReader.readXrefSection(PdfReader.java:839)
at com.itextpdf.kernel.pdf.PdfReader.readXref(PdfReader.java:777)
at com.itextpdf.kernel.pdf.PdfReader.readPdf(PdfReader.java:532)
at com.itextpdf.kernel.pdf.PdfDocument.open(PdfDocument.java:1638)
at com.itextpdf.kernel.pdf.PdfDocument.<init>(PdfDocument.java:231)
...
...
...
Caused by: com.itextpdf.io.IOException: file position {0} cross reference entry in this xref subsection.

The pdf used in the merge that's causing issues opens just fine in other programs. It is only itext7 that is complaining.

I can't share the problematic pdf, but I can say it's using PDF version 1.6.

This causes the result to be an empty pdf. What does this error mean and how can I fix it or work around it?

Answer 1

I found out that merger is actually working and merging the files but is throwing the exception anyway. I guess to inform the developer or user it is doing a fix? If you read the exception carefully it says the "Cross reference table will be rebuilt." So iText is doing some kind of automated fix.

This exception was triggering a different try catch block and the pdf was never being closed as a result.

Strangely I couldn't find any documentation on this in the iText 7 documentation.

At any rate, a simple try catch around the merge function fixed the issue in my case.

    for (InputStream in : files) {
    //Add pages from the each document
        PdfDocument sourcePdf = new PdfDocument(new PdfReader(in));

        try{
            merger.merge(sourcePdf, 1, sourcePdf.getNumberOfPages());
        } catch (Exception e) {
          //log an error and continue
        }
        sourcePdf.close();
    }