How to insert pages into a finished PDF

How to insert pages into an existing PDF

The contract is laid out, sent for approval, and then it turns out the annex is missing. Or the client sends a corrected page to replace one already in the document. You do not have to rebuild the file from scratch: inserting adds pages to an existing document without disturbing anything else.

Inserting differs from merging not in scale but in the fact that it has a base document. When you merge, all files are equal and stack up in sequence. When you insert, one file is the base and the others are embedded inside it at a chosen point. Several non-obvious things follow from that, and the first concerns the point.

Where pages land if you do not say

The answer surprises people: at the very beginning.

The default insert position is zero, and zero means the start of the document. The logic is formal rather than semantic: the tool does not try to guess that an annex usually belongs at the end. If you leave the position field untouched, the new pages land ahead of the title page.

So always state the point, even when it feels obvious where the annex should go.

Two ways to name the point, and which one wins

You can either give the page number to insert after, or an absolute position.

If both are set, "after page" wins and the position is simply ignored. Worth knowing: if you adjusted the position and nothing changed, a page-number field is probably still filled in somewhere.

Both values are clamped to the size of the document. Naming a page beyond what the file holds will not break the job, it just puts the insertion at the end.

Step by step

1. Open insert-pages and upload the base document first. 2. Add the files whose pages you want inserted. 3. Name the insertion point instead of trusting the default. 4. Run the job. 5. Open the result at the insertion boundaries: the page before, the inserted sheets, the page after.

If there are several files to insert, all of them land at the same point, one after another, in the order they occupy in the list. You cannot scatter them across different places in the document in a single pass; that takes several runs.

Rotation of inserted pages

Orientation alignment is on by default, and it does not work the way people expect.

The tool looks at the orientation of the first page of the base document and brings the inserted sheets into line with it. If an inserted page is landscape while the base document's first page is portrait, the inserted one gets rotated ninety degrees.

Orientation here is judged by how the page looks, not by the dimensions in its properties. A portrait sheet carrying a recorded ninety-degree rotation displays as landscape, and the displayed appearance is what counts.

A trap follows from that. If the base document opens with a landscape page, a cover or a fold-out diagram for instance, portrait annexes get turned to match it. Formally correct, semantically wrong. In that case switch the alignment off and sort out orientation with rotate-pdf.

What happens to bookmarks

Here inserting behaves the opposite way to merging, and that is deliberate.

Merging carries bookmarks over from every source file. Inserting keeps only the bookmarks of the base document and discards the navigation of the inserted files. The reason is simple: an inserted file's bookmarks point at page numbers inside that file, and in the combined document those numbers mean entirely different pages. Carrying them over would produce a table of contents that leads nowhere useful.

The base document's bookmarks survive intact: they are recalculated for the new page positions. A section that sat on page twenty and ended up on page twenty-five still works.

Separately, remember that page numbers printed inside the document itself are not recalculated by an insert. That is just text on a sheet. If the numbering has to run continuously afterwards, stamp it again with add-page-numbers.

Document properties come from the base file

The author, title and other properties of the result are inherited from the base document, and the properties of the inserted files are not mixed in. That is convenient behaviour: a package does not change its name in the properties just because someone added an annex.

The flip side is that if the base document's title was junk to begin with, inserting will not fix it. Correct that separately with set-pdf-metadata.

If the job will not start

There are only a few causes, and the message always names the specific file rather than complaining in general.

First: one of the files to insert is a PDF with no pages. That happens with documents that saved incorrectly or came out empty from some export. The tool checks this before starting and states plainly which file is empty.

Second: a file needs a password to open. Protection that only limits printing or copying does not interfere; a password on opening does. Clear it with unlock-pdf.

Third: the file is corrupted. Before declaring that, the service makes one attempt to repair the structure itself, so a message about an unreadable file means the automatic repair already failed.

A worked case: a signed sheet replacing the original

The most common task looks like this. A fourteen-page contract, page eight printed, signed and scanned. The scan has to take the place of the original page eight.

It does not happen in one operation, because inserting only adds. The sequence goes like this.

First remove-pages drops page eight. Thirteen sheets remain, and what used to be page nine is now page eight.

Then insert-pages puts the scan after page seven. After seven, not after eight: the numbering has already shifted, and page eight is now the former page nine. The off-by-one error is the most frequent one here, and people usually notice it after sending.

Check the scan's orientation separately. A scan of a signed sheet often arrives rotated, and if the contract's first page is portrait the automatic alignment will turn it for you. If the contract opens with a landscape cover, the alignment works against you, and the sheet is better prepared in advance.

You can reverse the order of operations, inserting first and removing after, but then the numbering shift has to be counted the other way. One rule is easier to remember: after every operation the page numbers change, and the next step is counted against the new document.

When inserting is not the right tool

Replacing a page in a single operation is impossible: inserting only adds. Replacement takes two steps, remove-pages for the outdated sheet and insert-pages for the new one.

To swap sections around inside a document, use organize-pdf. If the source files are equal and simply run in sequence, merging with merge-pdf is simpler and more honest: it preserves the navigation of every part, not just the first.