How to open a large CSV file (over 1 million rows)

An Excel worksheet holds 1,048,576 rows by 16,384 columns, and LibreOffice Calc has the same row limit. When a CSV has more rows, Excel loads what fits and warns: "This data set is too large for the Excel grid. If you save this workbook, you'll lose data that wasn't loaded." Don't save over the original file, or the missing rows are gone.

Google Sheets stops at 20 million cells or 100 MB per spreadsheet. A CSV with 20 columns reaches 20 million cells at 1,000,000 rows, often before it hits the file size limit.

rows.page reads the file with DuckDB-WASM inside your browser tab, so nothing is uploaded and you don't need an account.

1. Open https://rows.page and drop your .csv, .tsv or .txt file on the page, or choose it from your disk.
2. Scroll the grid. There is no row limit; rows load in pages of 100 as you scroll.
3. Open the Columns panel to see each column's type, distinct count, share of empty values, min and max, and a histogram or its top 10 values.
4. Filter in the search box, for example country:NL, price:>20 or -email:* for rows with no email. For anything more, type SQL that starts with SELECT, WITH or FROM and query the table data, such as SELECT country, count(*) FROM data GROUP BY country.
5. Open the Export menu and choose CSV. A filtered view exports every matching row; a SQL result exports up to 1,000,000 rows.

A browser tab holds up to 1.5 GB of CSV or TSV. If your file is bigger, rows.page loads the first part and shows "Loaded the first N rows: the file is X, more than the browser can hold." To work with every row of such a file, use DuckDB on your computer.

The DuckDB command-line tool reads CSV files of any size straight from disk. Start with an exact row count:

duckdb -c "SELECT count(*) FROM 'big.csv'"

To explore the whole file in rows.page afterwards, convert it to Parquet once. Parquet is compressed and typed, and rows.page doesn't hold Parquet files to the 1.5 GB check.

duckdb -c "COPY (SELECT * FROM 'big.csv') TO 'big.parquet' (FORMAT parquet, COMPRESSION zstd)"

If you only need to read or search the raw text, EmEditor 64-bit opens files up to 16 TB and costs $60 per year.

If you have to stay in Excel for Windows, load the CSV into the Data Model with Power Query. The rows don't appear in the worksheet grid, but PivotTables can use all of them.

1. In a blank workbook, go to Data > Get Data > From File > From Text/CSV and pick the file.
2. In the preview window, click the arrow next to Load and choose Load To.
3. Select Only Create Connection, tick Add this data to the Data Model, and click OK.

Microsoft's own advice for files that overflow the grid is the same import with Load To > PivotTable Report, which summarizes the full data set in a PivotTable.

To hand the data to someone who works in a spreadsheet, export it in pieces under 1,048,576 rows. In rows.page, filter one category at a time, such as country:NL, and export each view. Or run this SQL and raise OFFSET by 1,000,000 for every next part:

SELECT * FROM data ORDER BY __rid LIMIT 1000000 OFFSET 0

__rid is a hidden row number, so ORDER BY __rid keeps the original file order. It isn't included in the exported file.