"Error: Expected instance of PDFDict or PDFStream, but got instance of undefined" on save()

I have a funky (possibly corrupt) PDF that fails to fill. When I run save(), I hit the following error:

Error: Expected instance of PDFDict or PDFStream, but got instance of undefined
    at new UnexpectedObjectTypeError (/path/to/myproject/node_modules/pdf-lib/cjs/core/errors.js:40:24)
    at PDFContext.lookup (/path/to/myproject/node_modules/pdf-lib/cjs/core/PDFContext.js:78:15)
    at PDFDict.lookup (/path/to/myproject/node_modules/pdf-lib/cjs/core/objects/PDFDict.js:51:43)
    at PDFWidgetAnnotation.PDFAnnotation.getAppearances (/path/to/myproject/node_modules/pdf-lib/cjs/core/annotation/PDFAnnotation.js:75:20)
    at PDFTextField.needsAppearancesUpdate (/path/to/myproject/node_modules/pdf-lib/cjs/api/form/PDFTextField.js:619:48)
    at PDFForm.updateFieldAppearances (/path/to/myproject/node_modules/pdf-lib/cjs/api/form/PDFForm.js:460:23)
    at PDFDocument.<anonymous> (/path/to/myproject/node_modules/pdf-lib/cjs/api/PDFDocument.js:1150:38)
    at step (/path/to/myproject/node_modules/tslib/tslib.js:141:27)
    at Object.next (/path/to/myproject/node_modules/tslib/tslib.js:122:57)
    at /path/to/myproject/node_modules/tslib/tslib.js:115:75

I was able to fix this locally by adding a try{}catch{} around needApperancesUpdate. Should there be some way to fail gracefully here? Or any idea why the error might be happening?

The PDF: CCC901.pdf

[aman-saraf]

@schlosser . I don’t think that adding try catch block locally is a good idea. From my investigation, it looks like that the PDF document is corrupt. It seems that the pdf-lib expects the PDF in some standards, even though the load method won’t throw error for a corrupt pdf file.

In order to check whether your pdf file is compliant with standards, you may use online checkers such as https://www.pdfen.com/pdf-a-validator . Most probably, it will notify you of the errors & you may convert the pdf into specific standard & try again .

I hope it helps !

Update

I just tried updateFieldAppearances on save for a corrupt PDF file and it worked like a charm.

You may use

await pdfDoc.save({updateFieldAppearances: false})

and see if error still appears.

Update

The updateFieldAppearances doesn’t seem to work for multiple fields or it may also obscure the updated text which will be only visible once you click on the field.

Hence, the conversion of standard from PDF to PDF/A seems to be the only solution for the above problem.

[aman-saraf]

@Wian-TMC . Glad to know that it worked for you. Yeah, you’re correct in pointing the flaw in updateFieldAppearances. The conversion of PDF standard to PDF/A standard seems to be the best solution.