Why can’t you calculate with the numbers after converting an image to Excel?
Learn how to check OCR tables for text numbers, leading zeros, ambiguous dates and misplaced cells before converting data types and rebuilding formulas.
Converting a table image to Excel saves retyping. Whether the result is ready for calculations depends on column alignment, how the numbers have been interpreted and the data type stored in each cell.
Even a tidy worksheet can contain numbers stored as text, misplaced decimal points and missing leading zeros. These issues may be unobtrusive when reading, yet affect sorting, filtering and totals.
Recognizing digits is not the same as producing numeric values
An image records what characters look like, not spreadsheet data types. OCR may recognize 125.50, but those characters alone do not establish whether it is an amount, an identifier or a version number.
DocCrunch Image to Excel recognizes text in images and arranges it in a worksheet. Currently, exported cells are stored as text, including entries that look like numbers.
Being able to open the file in Excel and edit its cells does not mean an entire column is ready to sum. Some operations convert text numbers; some functions may ignore them. Results can therefore differ.
First determine what each column represents, then convert the relevant data types. Amounts and quantities usually need numeric values; product codes, postal codes and order numbers should generally remain text.
Not every sequence of digits should become a number
Similar-looking entries can require different treatment:
| Recognized entry | Possible meaning | What to check |
|---|---|---|
001204 |
Product or record identifier | Preserve leading zeros, usually by keeping text |
1,250.50 |
Amount | Establish the comma and decimal conventions before conversion |
03/04/2026 |
Date | Confirm whether the source uses month/day or day/month |
12% |
Proportion | Check that conversion represents 0.12, not 12 |
Separators need particular attention in tables from different languages or regions. 1,250 can mean one thousand two hundred and fifty or one point two five zero. A single cell is not enough to decide: consult column headings, units and the conventions elsewhere in the table.
Do not replace every blank with zero either. A blank may mean missing, not entered or not applicable; zero is an explicit value. Averages and counts can treat them differently.
Check rows and columns before calculations
OCR errors extend beyond individual characters. A correctly recognized number placed in the next row can be just as misleading.
Start by checking whether:
- Multirow headers have become part of the data.
- Long names have split into two records.
- Entries beside blank cells have shifted left or right.
- Subtotals, totals and footnotes have become ordinary records.
- Repeated page headers remain within a continuous table.
Review the name, quantity, unit price and amount together within each row. Checking that one column merely “looks numeric” is not enough.
If the image contains large irrelevant areas, use Image Crop to retain the complete table region first. Keep headings, units and necessary notes: a cleaner image is less useful if it loses the context needed to interpret the data.
Putting several images in one sheet does not finish consolidation
Photographed pages often have different column widths, repeated headers or explanatory text between tables. Exporting them into one file only brings the content together.
DocCrunch Image to Excel can create a separate sheet for each image or put all images into one sheet. The latter appends each image’s content in order, separated by filenames and blank rows. It does not automatically align column structures or remove repeated headers.
For images from one continuous table, check the column order, remove separator content and then organize a data range suitable for filtering and calculation.
If the columns have different meanings, separate worksheets are often easier to manage. Even two columns both labelled “Amount” need checking for matching currencies, units and scope.
A pictured result does not contain the original formula
If a screenshot shows 240.00, OCR can extract that result. It cannot tell whether the original cell used multiplication, a conditional formula or a reference to another sheet.
Rebuild formulas after verifying the data. Quantity times unit price, group subtotals and overall totals can help with checking, but matching totals do not prove every row is correct: two errors may cancel each other out.
Use the original spreadsheet if you have it. For a PDF with selectable text, try PDF to Excel before taking screenshots and running OCR. Scanned PDF pages are still images and require OCR; DocCrunch’s current PDF to Excel tool primarily extracts existing text and does not directly recognize scans.
Review the converted table in a consistent order
Before using the worksheet further:
- Compare it with the original image to confirm record counts and row and column alignment.
- Distinguish identifiers, text, numbers and dates, then handle types by column.
- Standardize decimal separators, units and date notation.
- Check blanks, minus signs, decimal points and repeated headers.
- Rebuild required formulas and cross-check against source subtotals or totals.
The value of Image to Excel is turning cell-by-cell typing into review and cleanup. Keep the source image as a reference and give the recognized data clear meaning before relying on the exported worksheet for calculations.
Recommended DocCrunch tools
These tools run locally in the browser whenever possible, so files do not need to be uploaded to a server.