How many columns can one extraction have?
Up to 30 columns in a single job. That is a ceiling rather than a target — a set you can check in one screen is easier to trust, and columns can be added on a later run once you have seen the first result.
Custom columns
Instead of accepting a fixed template, name the columns your spreadsheet needs, give each one a type, and get one row per document under exactly those headings.
Write down the headings you would put across the top of the sheet if you were typing it by hand. That list is the input — the documents are read against it, rather than the other way round.
A heading is an instruction. “Supplier” and “Supplier VAT Number” pull different values out of the same page, so name the fact you actually want in the cell.
Each column is text, number, currency or date. Where a heading could mean more than one thing on the page, a short description settles it before the batch runs rather than after.
The first result tells you whether the column set is right while it is still cheap to change. Adjust a heading, re-run, and only then send the whole folder through the same set.
A worked example
Product specification sheets from three manufacturers, each laid out differently and none of them a standard business document. There is no ready-made template for this and there never will be — the five headings below are simply the ones this buyer needed. The driver datasheet states no ingress rating, so that cell stays blank rather than being filled with a rating the document does not claim.
| Source PDF | Model | Input Voltage | Power Rating | IP Rating | Warranty |
|---|---|---|---|---|---|
| axis-lp40-datasheet.pdf | LP-40 | 220–240 V | 38 W | IP65 | 5 years |
| northvolt-driver-nd2100.pdf | ND-2100 | 100–277 V | 60 W | 3 years | |
| lumen-panel-600.pdf | LP-600 | 220–240 V | 40 W | IP20 | 5 years |
Tools that convert documents into spreadsheets usually ship a list of document types and a fixed field set for each. That works for as long as your paperwork looks like somebody else’s. The moment the job is specification sheets, inspection certificates, lab results, lease schedules or anything a supplier invented, there is no template to choose and no way to ask for the handful of things you actually need.
Naming the columns yourself removes that ceiling. The column set is the input to the job: whatever the documents look like, the spreadsheet comes back with your headings across the top and one row per file underneath.
A column heading is closer to an instruction than to a label, and small differences in wording change what lands in the cell. These five habits do most of the work.
Extracting specific fields from PDFs works through the same problem at more length, including what to do when two documents use different words for the same fact.
Every column carries a type — text, number, currency or date — and a type is what makes the finished workbook behave like a spreadsheet rather than a wall of text. A date column sorts; a currency column totals; a reference that merely looks numeric belongs in a text column so it keeps its leading zeros.
A description is optional and is worth writing whenever a heading is ambiguous on the page in front of you. One extraction can carry up to 30 columns, which is a ceiling rather than a suggestion: a set small enough to check on one screen is a set you can trust, and nothing stops you adding a column on a later run once you have seen how the first one came back.
Some documents are standard enough to be worth a starting point, so a few pages open the converter with columns already filled in: receipt to Excel and purchase order to Excel both do. They are starting points, not fixed templates — rename a heading, change a type, delete what your documents do not carry, and add what they do. If the job is invoices, invoice to Excel shows a worked accounts-payable example, and PDF to Excel covers what this workflow does and does not try to do with a page.
Once a set has produced a spreadsheet you were happy with, it is worth keeping. With a free account you can save the setup after a conversion and pick it again under Choose extraction setup, so the next batch starts at the upload step. Saving is the only thing that stores a setup — opening one of the pages above never adds anything to your account.
When the whole point is running one set across a folder, multiple PDFs to one Excel covers the batch side, and how it works walks through the review step before the workbook is downloaded.
Some of the work around a column set needs no account. Split PDF separates a combined file into the documents that should each become a row, Merge PDF joins pages that belong to one record, and Extract Text shows you the words a document actually contains, which is a quick way to see whether the heading you had in mind appears in it at all. The free tools page lists the rest.
There is no subscription — you buy credits and spend them when you convert something. The pricing page has the current packs and what a credit covers, and the FAQ covers batch limits, retention, refunds and failed files.
Type the headings, upload the documents, and check each row beside the PDF it came from before you download the workbook.
Choose your own columnsUp to 30 columns in a single job. That is a ceiling rather than a target — a set you can check in one screen is easier to trust, and columns can be added on a later run once you have seen the first result.
Name the single fact you want in the cell, using the wording the documents themselves use. A heading that names an area of the page rather than a value has nothing definite to return, and a heading that combines two facts produces a cell you then have to split.
Each column is one of text, number, currency or date. Typing a date or an amount is what makes the finished sheet sort and total without cleaning it first; text is the right answer for a reference that only looks like a number.
Yes. The receipt and purchase-order pages open the converter with a starting set of columns for those documents, and every column on them can be renamed, retyped, removed or added to before you upload anything. They are a starting point, not a fixed template.
Yes. After a conversion you can save the column set with a free account and choose it again under “Choose extraction setup”, which skips describing the fields a second time. Saving is the only way a setup is stored — nothing is added to your account just for opening a page.
The cell is left blank. A visible gap is a fact about that document, and it is more useful than a plausible value that quietly makes the spreadsheet wrong in a way nobody can see.
Yes, and that is the point of naming the columns yourself. A field can sit in a different place, or under a different label, in each document and still land under the heading you chose, because the column set describes what you want rather than where it sits on the page.