Excel throws an error, your text editor hangs, and the file just won't open. A genuinely large CSV — tens of millions of rows, hundreds of megabytes — outgrows the usual tools. Here's why, and how to actually work with it.
Most tools load the entire file into memory before showing you anything, so a 500 MB CSV needs that much RAM (often several times more once parsed) just to display. Spreadsheets add a hard row cap on top of that, and plain text editors try to render every line at once.
So a file that is perfectly valid simply exceeds what a desktop app will load — the problem is the tool, not the data.
Excel stops at 1,048,576 rows and 16,384 columns — and a real export can blow past that easily. Google Sheets caps at 10 million cells total (so a 20-column sheet maxes out around 500k rows). Apple Numbers is lower still. A few-hundred-MB CSV is routinely tens of millions of rows, well beyond all of them.
Hitting the cap is silent-ish: Excel imports the first ~1M rows and quietly drops the rest, so you can end up analyzing a truncated file without realizing it.
You can chop a big CSV into smaller files by row count or by a key column. It works, but it fragments the data: any analysis that spans the whole set (totals, dedupe, joins) now has to be stitched back together, and it's easy to lose a header or miscount a boundary.
Split when you genuinely need smaller files (a row cap downstream, an upload limit). Otherwise, prefer a tool that just handles the whole file.
This site parses big CSVs off the main thread (in a Web Worker) and virtualizes the table — it only renders the rows on screen — so the tab stays responsive on files far past Excel's row limit, with a live progress bar as it reads. You get a sortable table, KPIs and charts without the whole thing freezing.
And because it all runs locally, the file is never uploaded — a multi-hundred-MB export stays entirely on your machine.
When you need aggregates on a large file — GROUP BY, JOIN, filters — run SQL against it directly. The query tool can load a WebAssembly database engine in your browser for heavier work, so you get database-grade queries on a big CSV with nothing leaving your device.
That's usually faster and safer than coaxing a giant file into a spreadsheet just to pivot it.
Comfortably into the hundreds of MB and millions of rows on a modern machine — far past Excel's 1,048,576-row cap. The ceiling is your device's memory, not a fixed app limit, and the table is virtualized so rendering stays smooth.
No. Parsing, filtering, charting and SQL all run in your browser. You can drop a confidential multi-hundred-MB export and watch the Network tab stay silent.
Only if something downstream truly needs smaller files. Splitting fragments any whole-dataset analysis, so if you just want to view, filter or aggregate, open the whole file in the browser instead.
It silently truncates at 1,048,576 rows — anything past that is dropped on import. If your file is larger, Excel is showing you an incomplete picture.