How to get tables from a page or PDF
By Sure INSTANT team · Updated on Nov 12, 2026
A table on a web page or in a PDF report looks like data, but copying it into a spreadsheet breaks the columns and turns numbers into text.
This guide explains what makes a table easy or hard to read, how merged cells and headers are handled, and how SmartTable Capture gives you Excel, CSV and JSON.
Which sources work
| Source | Can its tables be read? |
|---|---|
| A public web page | Yes, every HTML table on it |
| A PDF made from a document | Yes, when you can select its text |
| A scanned PDF or a photo of a page | No: it is a picture without text |
| A saved HTML file | Yes |
Merged cells and header rows
Many tables merge cells across columns or rows, and some have two header rows. In Excel the merges are kept and the headers are in bold and frozen.
In CSV and JSON a merged value is repeated in every cell it covers, so each row is complete, and two header rows are joined into one column name.
Numbers, dates and Arabic digits
- Thousands separators, decimals, percentages and negative numbers in brackets become numbers.
- Common date forms become real dates in Excel and year-month-day in CSV and JSON.
- Arabic-Indic digits stay as written unless you switch on normalisation, which turns them into numbers.
Excel, CSV or JSON
- Excel: one sheet per table, for reading, sorting and calculating.
- CSV: one plain file per table, for importing into another program.
- JSON: every row with its column names, for developers and scripts.
How to do it with SmartTable Capture
- Paste the link of the page, or add the PDF or HTML file.
- Look at the free list of tables with their first rows.
- Keep the tables you need.
- Run it to download Excel, CSV and JSON.