TL;DR:
Moving data is three jobs: read it correctly, map each piece to the right column, and check that the totals agree.
The total on the source is your best test, because a typo almost always breaks it.
Keep the original open and review the result before you rely on it.
You have a supplier's order as a PDF and you want it in a spreadsheet so you can add it to your records. Retyping is slow and tedious, and tedium is how typos happen. This is the kind of task an assistant can do well, as long as it has a way to prove it did it right.
Here is a small version of the job, with the check that matters most.
The source and the target
The PDF has a table with four lines. The sheet has columns for Item, Quantity, Unit price and Line total.
Item | Qty | Unit price | Line total |
|---|---|---|---|
Pens | 12 | 1.50 | 18.00 |
Notebooks | 5 | 4.00 | 20.00 |
Tape | 3 | 2.25 | 6.75 |
Folders | 10 | 0.80 | 8.00 |
Walk-through: copying rows from a PDF into a sheet
Read the rows: It reads each line of the table, taking the item name and the three numbers. It keeps the decimals exactly as shown.
Map the fields: Item goes in column A, Qty in B, Unit price in C, Line total in D. It checks the header names against the sheet headers, not just the positions.
Enter the values: It types four rows. Numbers go in as numbers, not text, so the sheet can add them.
Check each line: For every row it multiplies:
12 × 1.50 = 18.00,5 × 4.00 = 20.00,3 × 2.25 = 6.75,10 × 0.80 = 8.00. All four match the line totals.Check the grand total:
18.00 + 20.00 + 6.75 + 8.00 = 52.75, which equals the total printed on the PDF. The data agrees with the source.
Two independent checks passed: each line multiplies out, and the lines add up to the stated total. If a single digit had been misread, such as 2.25 read as 2.28, the line check would have failed.
Notice what the assistant did not do. It did not trust its own reading. It treated the multiplication as a test and the printed total as an answer key. That is the mindset to look for: any tool that moves data should tell you how it checked its own work, not just that it finished.
A useful extra is a small "check" cell in the sheet, such as a formula that adds the line totals and compares them to the figure you typed from the PDF. If the two ever differ, the cell shows it at once, and the check stays in the sheet for whoever opens it next.
What goes wrong in this job
A digit misread from the PDF, especially in scanned documents.
A row skipped or doubled when a table runs over a page.
Numbers saved as text, so the sheet's total does not add them.
Columns shifted, so a price lands in the quantity column.
Currency symbols or thousand separators that confuse the sheet.
All of these are caught by the same two habits: compare each line against the source, and compare the grand total.
Mapping by header, not by position
Columns in a source can be in a different order than your sheet. If the PDF lists Item, Unit price, Qty and your sheet wants Item, Qty, Unit price, copying by position puts prices in the quantity column. Mapping by header name avoids this. If a header is unclear, such as "Amt", a careful assistant asks which of your columns it belongs in instead of choosing.
Keep the original in view
Leave the PDF open beside the sheet and scan down both when it is done. For a large table, check a sample from the top, middle and bottom, plus the total. Mistakes cluster at page breaks and in unusual rows, so look there first.
Tip: If the source has no printed total, make one: add the line totals in the PDF with a calculator, and compare it to the sheet's sum.
What this does not prove
Matching totals show the numbers are consistent, not that they are in the right columns. An item name could land on the wrong row and the totals would still match. A quick look at the names completes the check.
Privacy and control
Business documents can be confidential, so make sure you may share them. Working in your apps is among DuckyHelper's agent features, which are early access, and it's built to ask first before it sends a message or pays for anything.
Try it yourself
Take any small table, a receipt works well, and copy it into a sheet by hand. Check each line by multiplying, then check the total against the receipt. It takes two minutes and shows why the total is the first thing to test.
Frequently asked questions
Why not copy and paste from the PDF?
You can, but the table often pastes badly, with merged cells or text. You still need to check the result.
What if the PDF is a scan?
Reading from images is less exact. Check numbers more carefully.
How many rows can it do?
Any number, but longer tables mean more checking. Test a small batch first.
Get started
