Excel row limit: 1,048,576 rows, and what to do when you hit it
The Excel row limit is 1,048,576 rows and 16,384 columns per worksheet. It cannot be increased. No setting, no add-in and no version of Excel raises it, including Microsoft 365 and Excel 2024.
If your CSV has more rows than that, Excel opens the first 1,048,576 and shows “File not loaded completely”. The rest of your data is still in the file on disk. It is just not on your screen.
Why 1,048,576
It is 2 to the power of 20. Excel addresses rows with a 20-bit index, so the ceiling falls out of the file format rather than being a policy choice. The column limit, 16,384, is 2 to the power of 14, which is why columns stop at XFD.
The limit went up once, in Excel 2007, from 65,536 rows. It has not moved since.
The CSV row limit is different
A CSV file has no row limit. It is a text file, and text files have no structural maximum. A 40 million row CSV is a perfectly valid CSV.
The limit belongs to the program opening it. That distinction matters: your data is not damaged and nothing has been lost. You need a different way to look at it.
Can the Excel row limit be increased?
No. The workarounds people suggest each solve a different problem:
| Suggestion | What it actually does |
|---|---|
| 64-bit Excel | Raises the memory ceiling, not the row limit. Helps with crashes, not with row counts. |
| Power Query | Loads into the data model, which has no row limit, but you cannot browse or edit the rows. |
| Power Pivot | The same, aggregation only. |
| Splitting the file | Works, at the cost of every cross-file total being manual. |
| Registry hacks to raise the limit | These do not exist. Any page offering one is wrong. |
Five ways to work past it
1. Count the rows first. On macOS or Linux:
wc -l bigfile.csvOn Windows PowerShell:
(Get-Content bigfile.csv -ReadCount 1000 | Measure-Object -Line).LinesUnder a million means your problem is memory or column count, and the fix is a different one.
2. Split the file. split -l 1000000 bigfile.csv part_ on macOS or Linux gives you chunks Excel opens. The cost is real: every sum, every pivot and every filter now has to be done nine times and reconciled by hand. That is where the errors come from.
3. Power Query. Data, then Get Data, then From Text/CSV, then Load To, and choose “Only Create Connection” with “Add this data to the Data Model”. The row limit does not apply because nothing lands in the grid. Good for a repeatable summary, poor for looking at your data or editing a cell.
4. Load it into a database. DuckDB reads a CSV in place with no import step:
duckdb -c "SELECT count(*) FROM 'bigfile.csv'"The right answer if you write SQL. Not an answer if you have a report due on Thursday and you do not.
5. Use a spreadsheet with no row limit. Some desktop tools never load the whole file. They index it, read only the rows on screen, and push sorting and filtering down to a query engine. The row count stops being a number you have to think about.
The truncation trap
The dangerous failure is not the error message. It is this sequence:
- You open a 3 million row export.
- Excel warns that the file was not loaded completely. You click OK.
- You sum a column, build a pivot, and send the numbers on.
Those numbers were computed on the first third of your data, and nothing on screen says so. If you have ever clicked past that dialog and cannot remember what happened next, check the row count against the source file before trusting the result.
What we built
BigXL is a desktop spreadsheet with no row limit. It opens a 10 GB CSV in seconds because it never loads the whole thing into memory: rows appear as you scroll, and sorting, filtering and pivots run through an embedded query engine over the entire file.
It runs on your own machine with no account and no network calls, so a file with payroll or customer records in it never leaves the building.
Opening, sorting, filtering, charting and CSV export are free at any file size.