Documentation/ Uploading documents/ Multi-page PDFs and auto-merge
Multi-page PDFs and auto-merge
Every PDF is processed page by page, which is what makes the extraction reliable, and then auto-merge puts the pages of one document back together. It comes switched on, and there is only one case where you want it off.
Document upload › Auto-merge all
On this page
Why the reading goes page by page
It looks like the long way round, but it is what makes the reading dependable: processing each page on its own means none of them is left half done, however many the file has and however it is laid out.
The price of that decision is that a five-page document would turn into five documents. Which is where auto-merge comes in.
What auto-merge does
The screen sums it up itself: with auto-merge on, which is the default, invoices and delivery notes running to several pages are put back together on their own, with nothing for you to do.
There are two switches:
| Switch | What it covers |
|---|---|
| Auto-merge all, at the top | The whole batch. |
| Auto-merge, on each file | That one file. |
Which lets you do the usual thing: leave the general one on and turn it off only on the file that needs it.
The row while the pages are being merged
A multi-page PDF is read page by page, so the merge takes a few seconds. During that time the list shows a single row for the whole file, on a light blue background, with the file name, a spinning icon and the progress:
The label says how many pages have arrived: Merging pages · 3 of 5.
That row cannot be opened or ticked, and it does not count towards the totals in the header, because its data is still incomplete.
The list updates itself
While a row is being merged, the list refreshes itself every few seconds. There is no need to reload the page. Once the pages are merged, the row becomes the invoice or the delivery note, with all its data.
The refresh keeps your place: the page of the list, the scroll position and any documents you have ticked.
Pages that do not arrive
If five minutes go by with no new pages and the group is incomplete, the row turns orange and shows Missing pages · 3 of 5 with the text "Not all pages have arrived". The Merge received pages button merges the pages that did arrive into one document.
Seeing the pages separately
The Show pages button opens out the separate pages, so you can check one without waiting for the merge to finish. Each one carries a blue label beside its number — Page 2 of 5 — and clicking it groups them back into a single row.
The list keeps refreshing while you look at them separately. If the merge finishes meanwhile, the merged document appears straight away.
After 24 hours, a group that has not finished merging stops being shown as a group and its pages go back to the list as separate documents. From there they are merged by hand with Merge selected.
Lists where it appears
On the invoice, expense, sales and delivery note lists, and only on uploads with auto-merge on and files of more than one page. Groups of photos uploaded as a single document behave the same way.
The one case for turning it off
Turn it off when one PDF holds several different invoices. That keeps each page as a document of its own.
This is the PDF that arrives from the bookkeeper with twelve invoices one after another, or the scan of a whole batch. With auto-merge on, those twelve would become one.
Outside that case, leave it on. It is what saves you the manual work.
Dropping pages before you process
Click on a file and you see its pages, and you can take out the ones that carry nothing: covers, advertising, terms and conditions, blank sheets.
The screen reminds you why: you pay by the page processed. Taking three sheets of advertising out of every invoice from a regular supplier shows up in the month's usage.
If you forgot
You do not have to delete and start again. From the list:
- Merge selected joins exactly the ones you tick, at least two of them. Before it goes ahead, the dialog lets you take out any ticked by mistake.
If a PDF holding several invoices got merged by mistake, the way out is to split the document and build it again.
The totals after a merge
When documents are merged, the one you end up with keeps the totals and taxes it had at that moment: they are not worked out again. After a merge, or after splitting and re-merging, go over the header.
That is the check that saves you from the most annoying case of all: a document correctly put back together but carrying the totals of just one of its pages.
Things that come up
| What you see | Why | What to do |
|---|---|---|
| A PDF of twelve invoices came in as one | Auto-merge was on. | Split the document, and turn it off on that file next time. |
| A five-page invoice came in as five documents | Auto-merge was off. | Merge them from the list. |
| Usage is higher than you expected | Pages with nothing on them were processed. | Take them out before processing. |
| The totals do not add up after a merge | They are not recalculated when documents are joined. | Go over the header of the document you ended up with. |
| The blue row will not open | The pages are still being merged and the data is incomplete. | Wait for it to finish, or press Show pages to check one on its own. |
| The row has been orange for a while | Five minutes have passed without the rest of the pages arriving. | Press Merge received pages, or upload the whole file again. |
| You look for the Auto-merge button on the list | It was taken off: the automatic merge is decided at upload, with the tick box. | Merge by hand with Merge selected, or upload again with the box ticked. |