Delete PDF pages

The pages you name are absent from the file that comes out — copied out of existence rather than marked invisible. There is nothing left in the document to recover, which is the behaviour you want when the reason for removing a page is that it should not be seen.

The usual caveat still applies and is worth repeating: this is not redaction. A page you keep keeps everything on it, including text hidden under a black rectangle and content cropped out of view. Removing a page removes a page.

Removing pages 4 and 9 from a fifty-page document and keeping pages 4 and 9 out of it are opposite descriptions of the same kind of operation, and both tools exist because which one is less work depends entirely on your document. Naming two pages to drop is easier than naming forty-eight to keep.

So use this page when most of the document stays, and the extract page when most of it goes. Neither can do something the other cannot; they just spare you writing out the longer list and getting one number wrong in it.

The awkward part of this job is usually not the removal but the counting: finding out that the duplicate scan is page 23 and not page 22 means paging through the document first. The page count is shown once the file is read, and a number beyond it is refused with the real length rather than accepted and ignored.

A habit worth having with scans in particular: check the result rather than the input. A feeder that pulled two sheets at once produces a document where the page you meant to delete and the page after it are both wrong, and that is far easier to see in the output than to predict from the original.

Removing a quarter of the pages does not usually remove a quarter of the bytes. Two reasons: a PDF shares resources between pages, so a font or a logo used throughout is stored once and stays as long as any page still needs it — and the pages you delete are often the light ones, a blank separator or a cover, while the scanned photograph you kept is most of the file.

If size is the actual goal rather than a side effect, deleting pages is the wrong lever. Compressing the PDF works on what the remaining pages contain, which is where the weight really is.

Bookmarks and the document outline point at pages by position, and the positions after a deletion are not the ones they were written against. A table of contents that was generated as text is unaffected and now wrong — it still lists the page you removed, because it is ink on a page rather than a live reference.

Links between two pages that both survived keep working. Links to a page that went do not. None of this is recoverable afterwards from the result, so if the outline matters, do the deletion in an editor that rebuilds it.

Deleting several pages by index is a classic way to remove the wrong ones: take out page 2, and every page after it moves up, so the next deletion lands one place off. Tools that get this wrong produce a document that looks plausible and is missing the wrong sheet.

This one builds the result from the pages that stay rather than by removing the ones that go, so the numbers you typed mean what they meant when you typed them. And a selection that would remove every page is refused, because a PDF with no pages is not a document.