How to split a large PDF, and how to choose where to cut

A single PDF that has grown too big to work with is a familiar problem. A scanned bundle of two hundred pages, a report with an appendix nobody needs, a portal that will not accept anything over ten megabytes. Splitting is often the answer, but not always, and picking correctly saves a lot of repeated attempts.

Splitting or compressing?

The two solve different problems and are frequently confused.

Compression reduces file size while keeping every page. It is the right choice when the whole document must stay together and the only issue is bytes — a contract that is too large to email, a submission that must be one file.

Splitting divides the document into several smaller files. It is the right choice when the recipient only needs part of it, when different sections go to different people, or when a portal caps the size of each upload but allows several.

• Whole document needed, file too large: compress
• Recipient needs one section: split and send that section
• Portal caps each file but accepts several: split
• Confidential appendix that should not be shared: split and send only what is appropriate
• Still too large after compression: split as well

That last combination is worth remembering. Compressing first and then splitting produces smaller parts than splitting a bloated original, so if you are doing both, compress first.

Deciding where to cut

The mechanics are trivial. The judgement is in choosing boundaries that make the parts useful rather than arbitrary.

Cut at natural divisions in the document: chapters, sections, the start of an appendix, the boundary between a report and its supporting evidence. A recipient who receives part two of four should be able to tell what it contains from the first page.

Avoid cutting mid-table or mid-argument. A table split across two files becomes considerably harder to read than one that is slightly too long, and a page that begins halfway through a sentence signals carelessness.

Name the parts so the order is obvious and they sort correctly. Numbering with a leading zero matters more than people expect: without it, part ten sorts before part two in most file listings.

Working around upload limits

1. Find the actual limit. Portals often state one figure and enforce another, and the stated number is usually per file rather than per submission.
2. Compress the document first and check whether it now fits as a single file.
3. If it does not, split at section boundaries rather than at an even page count.
4. Aim comfortably below the limit rather than just under it, since some systems apply the cap after encoding, which inflates the size.
5. Test with the first part before preparing all of them, in case the portal rejects the format for an unrelated reason.

That last step has saved a great deal of wasted effort. Discovering that a portal refuses your file for some reason other than size is much cheaper on the first part than after you have prepared eight.

What splitting preserves and what it loses

The pages themselves come through unchanged: text stays selectable, images keep their resolution, and quality is unaffected, because splitting copies page objects rather than re-encoding them.

Document-level features do not always survive. Bookmarks pointing to pages that are now in another file cannot work. Internal cross-reference links break in the same way. A form spanning several pages may lose the relationship between its fields, and a digital signature covering the whole document will no longer validate, because the document it signed no longer exists in that form.

For a signed document this matters a great deal. Splitting a signed PDF produces parts that carry no valid signature, so if the signature is the point, send the whole file and compress instead.

Split when the recipient needs less than everything. Compress when they need everything and it will not fit.