PDFdesk

Blog · 2026-02-25 · 4 min

What merging quietly loses

Combining PDFs keeps every page and drops several things that are not pages. Knowing which ones saves rebuilding a packet twice.

Merging is the most-used PDF operation and one of the most reliable: page content, fonts and images all come across exactly as they were.

What comes across less reliably is everything that is not page content — and because the pages look right, the losses are easy to miss until someone else finds them.

What is usually lost

Bookmarks. The outline lives in the document catalogue, not on the pages. Merge two documents with careful sidebar structure and the result frequently has no outline at all. Not a broken one: absent, with no warning.

Form fields. Merging often drops the interactive layer, taking every filled-in answer with it. This is the expensive one, because the pages still show boxes where the answers were.

Internal links. A cross-reference pointing to page 12 of a document that is now page 84 of a packet either breaks or, worse, silently points at whatever is on page 12 now.

Attachments. Files embedded in a source document are not necessarily carried into the result.

Metadata. The output takes one document's properties, or none, and the rest is discarded. Usually harmless, occasionally not — a document whose title carried a matter reference, for instance.

The pattern behind it

All of these are document-level structures. Merging works page by page, and anything that belongs to the document rather than to a page has to be explicitly rebuilt for the combined result. Many tools do not, because the correct behaviour is often ambiguous — whose metadata wins? how should two outlines nest?

Silence is the wrong answer to an ambiguous question, but it is the easy one.

Working around it

Flatten forms before merging. Values become page content and survive.

Rebuild bookmarks afterwards, not before. Note where each source document starts *before* you merge — afterwards the boundaries are invisible, because a merged PDF has no idea where one document ended and the next began. Those start pages are exactly what an outline needs.

Check links if they matter. In a document set where cross-references are load-bearing, verify a sample after combining.

Set metadata deliberately on the finished packet rather than accepting whatever survived.

The order that works

Note the section boundaries. Flatten any forms. Merge. Add front matter. Add bookmarks last, because front matter rebuilds the document again and will discard an outline written before it.

None of this is difficult. All of it is invisible until someone opens your packet, clicks a bookmark that is not there, and asks.

Tools this is about