PDF/A error

PDF/A error: XMP metadata does not match the document information dictionary

When Title, Author or the dates disagree between XMP and the Info dictionary, PDF/A-1b fails. Why the duplication exists and how to reconcile it.

PDF/A requires the XMP metadata stream and the document information dictionary to agree wherever they describe the same property, and in this document they do not. Fix it by regenerating the XMP from the Info dictionary — or the reverse — so both carry the same values.

ISO 19005-1 clause 6.7Metadata

The Metadata check group covers 6.7.x XMP metadata, extension schemas, PDF/A identifier. The exact sub-clause is reported on the issue itself when the validator raises it.

What causes it

  • A tool updated the Info dictionary — a title, an author, the modification date — without regenerating the XMP packet.
  • Two tools in a pipeline each wrote one of the two representations.
  • The XMP packet was copied from a template and never reconciled with the actual document properties.
  • Date formats differ: the same instant written with a different timezone or precision in each place still counts as a mismatch.

How to fix it

  1. Regenerate the XMP packet from the document properties as the last step of the pipeline, after everything else has finished editing them.
  2. Convert the document: gopdfrab injects consistent metadata, including the PDF/A identifier, as part of conversion.
  3. If you write metadata yourself, write both representations from one source of truth rather than editing them independently.

Frequently asked

Why does a PDF store its metadata twice?

The Info dictionary is the original PDF mechanism; XMP is the later, extensible, XML-based one. PDF/A requires XMP and tolerates the Info dictionary, but insists the two never contradict each other — a reader must not get different answers depending on where it looks.

Does the PDF/A identifier live in XMP?

Yes. The part and conformance level are recorded in the XMP packet in the PDF/A identification schema. A file missing that identifier is not claiming conformance at all, whatever else is right about it.

Can I just delete the Info dictionary?

Removing the contradiction that way does resolve the mismatch, but leaves the document with less metadata than it had. Reconciling the two is usually better than discarding one.

Check your own document
The validator runs the same engine in your browser and names every failing check. The file is never uploaded.