The notation: every pattern you'll ever need
Page ranges are built from two kinds of tokens, joined by commas:
**Single page** — just the number: 5 → page 5 alone; 1 → the first page.
**Hyphenated range** — start and end, joined by a hyphen: 5-10 → pages 5, 6, 7, 8, 9, 10; 1-3 → pages 1, 2, 3.
Three rules that hold across virtually every tool:
**The single-page edge case** — To extract exactly one page, write the number alone: 5. Writing 5-5 also works in most tools and means the same thing — but the bare number is cleaner, and some strict parsers reject 5-5. (Illustrative: both forms select one page; prefer the bare number.)
**"From page N to the end"** — For "everything from page 10 onward," write 10-<last page> — e.g., 10-40 for a 40-page document. Some tools accept the open-ended 10- but many don't, so look up the total page count first (any PDF viewer shows it) and write the full range. A browser-based planner that detects the page count locally makes this painless: you always know the last page number.
- 1-3,7 → pages 1, 2, 3, and 7 (a range plus a single)
- 1-3,12-15 → pages 1–3 and 12–15 (two ranges)
- 1,3,5 → pages 1, 3, and 5 (scattered singles)
- 1-3,7,12-15 → pages 1–3, page 7, and pages 12–15 (everything mixed)
- Pages start at 1. The first page of the document is page 1, not page 0.
- Ranges are inclusive on both ends. 5-10 is six pages. This is the single most common mistake — people write 5-10 expecting five pages and get six. When in doubt, count: end minus start, plus one.
- A range can't run backward. 10-5 is invalid in nearly every tool; write 5-10.
The order caveat (read this before you extract)
Here's the surprise: some tools output pages in ascending order no matter what order you typed. You write 12-15,1-3 expecting pages 12, 13, 14, 15, 1, 2, 3 — and get 1, 2, 3, 12, 13, 14, 15.
This isn't a bug in most cases; it's a design choice (the tool treats your input as a set of pages, not a sequence). But it matters enormously when order matters:
Our planner sidesteps this cleanly: you define parts as explicit ranges, and the part list shows exactly which pages land in each part — "Part 1: pages 1–2 (2 pages)" — so there's no ambiguity about what goes where.
- Building a chapter from scattered pages? Type tokens in ascending order to begin with, and the caveat never bites.
- Reordering pages deliberately? You don't want an extractor at all — you want a merge/reorder tool. Extraction selects pages; it doesn't rearrange them.
- Unsure what your tool does? Test with a tiny range first: extract 3,1 from any PDF and check whether the output is page 3-then-1 or 1-then-3. Ten seconds, total clarity.
Three extraction workflows, start to finish
**Workflow 1: Pull one page (the receipt, the signature page, the invoice)** — Open the PDF and note the page number from the viewer's counter (not the printed footer — see the numbering trap below). In the planner at https://all-done.xyz/tools/split-pdf, type the single number: 7. Check the part list, download the .txt plan. Apply the plan in PDFsam Basic (Extract pages), Preview (Print page 7 → Save as PDF), or your online splitter. Total time: under two minutes, most of it finding the right page number.
**Workflow 2: Extract a chapter (pages 45–78 of a 200-page ebook)** — Type the range: 45-78. That's 34 pages (78 − 45 + 1 — count it). Verify the part list shows "pages 45–78 (34 pages)." Name the output now, not later: ebook-chapter-3.pdf beats extracted-pages.pdf in every future search. Apply in your cutting tool of choice. For multiple chapters at once, define them as separate tokens in one plan — 45-78,79-110,111-140 — and the planner lists each as its own part. One plan, three outputs, no repeated setup.
**Workflow 3: Odd/even duplex sets** — Duplex printers and print shops live on this: odds in one file, evens in another. Odds: 1,3,5,7,9 (keep going to the last odd page). Evens: 2,4,6,8,10 (keep going to the last even page). For a 20-page document this is quick to type. For a 200-page document, don't type it by hand — generate the list in a spreadsheet (odd numbers in one column, join with commas) or use a desktop tool with a real duplex mode. Some browser planners reject keyword tokens like "odd"/"even" outright (ours does — it only accepts explicit numbers and ranges), which is actually a feature: explicit numbers can't be misinterpreted.
Split vs delete: pick the right operation
People conflate these constantly, and picking the wrong one wastes real time:
Same result in theory, different effort in practice. Removing 3 bad pages from a 100-page document? Delete mode: type 23,57,88 and you're done. Extracting 3 good pages from a 100-page document? Extract mode: type 23,57,88 and you're done. The deciding factor is always which list is shorter — the pages you want, or the pages you don't.
- Extract/split = choose the pages you KEEP; they become a new file. Use it when you want a subset as its own document.
- Delete pages = choose the pages you REMOVE; the rest stays. Use it when you want the same document minus a few pages.
The numbering trap: viewer pages vs printed pages
The most common "wrong pages" complaint has nothing to do with syntax. A scanned book shows "page 1" in its footer on what the PDF viewer calls page 7 — because the PDF includes a cover, a title page, and four blank sheets the book's own numbering ignores.
Always use the PDF viewer's page counter, never the printed footer numbers, when writing ranges. If the viewer says you're on page 7 of 200, you write 7 — even if the footer says "page 1." When a document's internal numbering and printed numbering disagree, every splitter on earth follows the internal one.
Troubleshooting: when the range doesn't do what you expected
Range notation is simple, but four mistakes account for nearly every "it didn't work":
**"Invalid page range" errors.** The usual suspects: a backward range (10-5 — flip it to 5-10), page 0 (pages start at 1), a range past the last page (1-50 in a 40-page document), or stray characters (letters, semicolons, or words like "odd"/"even" that strict parsers reject). Fix the token, not the tool — the syntax rules above cover every valid case.
**Keyword tokens rejected.** Some people type odd, even, or all expecting the tool to understand. Most free planners don't — ours explicitly rejects anything that isn't a number or a hyphenated range. Spell the pages out (1,3,5,7) or use a desktop tool with a real duplex mode.
**Right syntax, wrong pages.** Re-read the numbering trap above: if the PDF's internal page order doesn't match the printed footers, your perfectly valid 45-78 extracts the wrong content. Verify with the viewer's counter, and when in doubt extract one test page first and check it.
**Empty or single-page output from a big range.** Almost always a reversed range or a typo (45-87 in a 78-page document silently clamps or fails depending on the tool). Count the expected pages — end minus start plus one — and compare with what came out.
From plan to file: the two-minute execution
Planning is the hard part; cutting is mechanical. Once your ranges are verified in the planner:
Keep the .txt plan alongside the outputs. Six months from now, when someone asks "which pages went into the appendix file?", the plan answers in seconds.
- PDFsam Basic (free, Windows/Mac/Linux): open the Extract module, enter your ranges exactly as planned, run. It handles multi-range plans in one pass.
- macOS Preview: open the PDF, File > Print, set Pages to your range, click PDF > Save as PDF. Repeat per part.
- Online splitter you trust: paste or select the same ranges; verify the output page counts match your plan.
