Skip to content

CSV too large for Excel? How to open every row

By MegaRows · Published · Updated · 9 min read

Excel stops at 1,048,576 rows. See every row here, free. Nothing is uploaded.

Open a CSV, Excel, TXT, JSON, MATLAB, Parquet, HDF5 or LabVIEW TDMS file

Drop it here or choose a file to chart it. Millions of rows are fine.

The sample is 1,000,000 hourly readings of made-up UK weather.

Smaller samples:

Tip: drop several files to compare them.

Your file never leaves your device. It is read and charted by your browser. Nothing is uploaded.

You double-click a CSV export, Excel thinks for a while, and then warns you that the file was not loaded completely. The file itself is fine: Excel simply cannot show more than 1,048,576 rows on a sheet. The tool above, the same as our large CSV viewer, opens all of it. This guide explains what happens to the rest of your data and walks through six more ways to work with the whole file, from tools you already have to ones that need a little code.

Why Excel stops at 1,048,576 rows

Since Excel 2007, every worksheet has exactly 1,048,576 rows (220) and 16,384 columns, up to column XFD. The older .xls format topped out at 65,536 rows. The limit is built into Excel’s grid and file format, so no setting, add-in or faster computer raises it. It is the same in Microsoft 365 on Windows and Mac and in Excel on the web.

Google Sheets counts cells instead of rows: a spreadsheet holds “up to 20 million cells”, as Google’s help page said when we checked it on 10 October 2026, so a file with 20 columns tops out at 1,000,000 rows.

The CSV format itself has no row limit. It is plain text, and a database export, web log or sensor recording can easily run to tens of millions of lines and several gigabytes. The problem is never the file: it is the program you open it with.

What you see when Excel truncates a CSV

When a CSV has more lines than the grid can hold, Excel loads the first 1,048,576 (header included) and shows a warning. Depending on your version, it says “File not loaded completely” or “This data set is too large for the Excel grid. If you save this workbook, you’ll lose data that wasn’t loaded.” Click OK and you get a normal-looking spreadsheet that is missing everything after row 1,048,576.

  • Check whether you were cut off. Click a cell in column A and press Ctrl+↓. If you land on row 1,048,576, the file was almost certainly truncated. Totals, counts and pivot tables then quietly cover only the loaded rows.
  • Don’t save over the original. If you save the truncated workbook as CSV under the same name, the missing rows are gone from disk for good. Keep the original export untouched and work on a copy.
  • Watch for silent conversions. Excel strips leading zeros from codes and phone numbers, turns long IDs into scientific notation (it keeps only 15 significant digits) and reinterprets dates using your regional settings. Recent Microsoft 365 versions let you switch some of this off under File › Options › Data › Automatic data conversion.

First, count the rows

Before choosing a tool, find out how big the problem is. A file with 1.2 million rows needs a different approach from one with 80 million. Choose Count rows in the tool at the top of this page and pick the file: you get the number of lines in seconds, without opening it in a spreadsheet. It is read on your device and never uploaded.

On the command line, wc -l data.csv (macOS and Linux) or find /c /v "" data.csv (Windows Command Prompt) gives the same answer. Remember that the count includes the header line, and that a value with a line break inside quotes spans more than one line. Our CSV row counter page explains the difference between lines and records.

Option 1: Split it into Excel-sized files

If you need the data in Excel itself, split the file into parts that each fit under the limit, with the header row repeated at the top of every part. At 1,000,000 data rows per file, a 3.4-million-row CSV becomes four files.

The quickest way is our CSV splitter: choose the file and click “Split & download”. It makes parts of 1,000,000 rows by default, repeats the header in every part, never cuts a quoted value with a line break in two, and gives you the parts as one ZIP. It runs in your browser, so nothing is uploaded, and it can also split by size or into one file per value of a column.

On macOS or Linux (or on Windows through WSL or Git Bash), the command line does it too:

# Keep the header, split the rest into parts of 1,000,000 lines
tail -n +2 data.csv | split -l 1000000 - part_
for f in part_*; do
  { head -n 1 data.csv; cat "$f"; } > "$f.csv" && rm "$f"
done

Pros: every part opens normally in any version of Excel; no new software for most people.

Cons: sorting, filtering and totals across parts are painful; a plain line-based split, like the commands above, can cut a record in half if values contain quoted line breaks.

Option 2: Power Query and the Data Model

Excel can work with far more rows than the grid shows, as long as you keep them out of the grid. Power Query loads the CSV into the Data Model, a compressed in-memory database inside the workbook, and PivotTables and PivotCharts then summarise all of the rows. In Excel for Windows:

  1. Go to Data › Get Data › From File › From Text/CSV and choose 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.
  4. Use Insert › PivotTable › From Data Model to summarise the full data set.

Power Query can also filter and clean the data before it loads. If you only need last year’s orders or a handful of columns, the result may fit in the grid after all. Power BI Desktop, a free Windows app, uses the same engine.

Pros: stays in Excel; handles tens of millions of rows, limited by memory; refreshes when the CSV changes.

Cons: you can’t scroll through the rows; the Data Model is only fully supported in Excel for Windows; big models make workbooks slow.

Option 3: Load it into a database

Databases are built for millions of rows. Two free options need no server and install in a minute: DuckDB, which queries CSV files directly, and SQLite, which imports them into a single database file.

-- DuckDB: query the CSV directly, no import step
SELECT count(*) FROM 'data.csv';
SELECT region, sum(amount) FROM 'data.csv' GROUP BY region;

-- SQLite (in the sqlite3 shell): import once, then query
.import --csv data.csv sales
SELECT count(*) FROM sales;

Graphical tools such as DBeaver or DB Browser for SQLite give you a spreadsheet-like view of query results.

Pros: very fast; copes with hundreds of millions of rows; results are often small enough to paste into Excel.

Cons: you need to install software and learn basic SQL; it is not a sheet you can scroll and edit freely.

Option 4: Python and pandas

If you are comfortable with a little code, pandas reads a CSV of a few million rows in seconds:

import pandas as pd

df = pd.read_csv("data.csv")
print(len(df))          # data rows, header excluded
print(df.head(20))      # first 20 rows
df[df["region"] == "North"].to_excel("north.xlsx", index=False)  # needs openpyxl

If the file is bigger than your memory, process it in chunks:

total = 0
for chunk in pd.read_csv("data.csv", chunksize=1_000_000):
    total += chunk["amount"].sum()

Polars is a faster alternative with a similar feel, and its lazy scan_csv handles files larger than memory.

Pros: free, very flexible and repeatable; ideal for cleaning, reshaping and automating.

Cons: needs Python and some coding; a DataFrame often uses several times the file size in RAM; no point-and-click grid.

Option 5: Command-line tools

For a quick look, the tools built into macOS and Linux (and available on Windows through WSL or Git Bash) are hard to beat:

head -n 20 data.csv          # first 20 lines
tail -n 20 data.csv          # last 20 lines
wc -l data.csv               # count lines
grep -c "London" data.csv    # count lines that contain a word

Dedicated CSV tools such as xsv, qsv and csvkit understand quoting and can select columns, sort, filter and compute statistics on files of many gigabytes.

Pros: instant, even on huge files; usually nothing to install; easy to script.

Cons: no grid; head and grep don’t understand quoted fields; less friendly on Windows.

Option 6: Browser-based tools, like the one above

Online CSV viewers need no installation, but check two things before dropping a large file into one. Viewers that upload your file are slow for big files and a problem for confidential data, and some cap the file size or row count.

With MegaRows everything runs locally in your browser, so the file never leaves your device. The tool at the top of this page opens the whole file on a table (the same engine as the large CSV viewer), counts its rows like the CSV row counter, or charts any column like the CSV plotter, which also opens Excel, JSON, MATLAB, Parquet, HDF5 and LabVIEW TDMS files. To get part of the data back into Excel, sort and filter the table to the rows you need, or zoom a chart into a range. Under “Do more”, “Download the table as shown (CSV)” saves the table’s rows as you see them, sorted and filtered, with every column; “Download all columns (CSV)” saves every column and every row, the whole file; and “Download chart data (CSV)” saves the plotted columns for the rows in view. Several exports to combine?Merge them into one CSV first.

A CSV opened with “Open and view”: the 1,000,000-row UK weather sample on a table of numbered rows, with each column’s type and unit in its header and the column list on the left.
“Open and view” with the 1,000,000-row UK weather sample: every row on a table.

Want a file to download and try first? The large CSV examples for testing go up to 100,000 rows (4.69 MB), well within Excel’s limit.

There is no fixed row limit: the practical limit is your device’s memory. In our tests a 2-million-row CSV opened in about 1 second, and a 10-million-row CSV opened and drew its first chart in about 3 seconds.

Measured in Chrome on a recent laptop (Apple M4): a 423 MB CSV with 10,000,000 rows and 5 columns, from choosing the file to the first chart. Slower devices take longer.

MegaRows also splits a CSV into Excel-sized files, turns an Excel workbook into a CSV with Excel to CSV and saves a CSV as a MATLAB file with CSV to MAT, all in your browser. Comparing two versions of a CSV by key is coming soon. Our guides also cover how to plot a large CSV and merge CSV files.

Pros: nothing to install, works on any operating system; local tools keep your data private.

Cons: limited by your browser’s memory; some online tools upload data or cap sizes.

Which option should you choose?

Comparison of ways to open a CSV file that is too big for Excel
OptionBest forInstallSee every row?Skill
Split the fileGetting the data into plain ExcelNone (our splitter, or the command line)Yes, part by partLow
Power QueryPivots and summaries in ExcelNone (Excel for Windows)NoMedium
DuckDB or SQLiteFiltering and aggregatingSmall downloadQuery resultsMedium (SQL)
Python and pandasCleaning and automationPythonIn codeHigh
Command lineA quick look, countingBuilt in (macOS, Linux)First and last linesMedium
Browser toolsQuick checks without installingNoneDepends on the toolLow

If you only need to know how big the file is, count it first. If colleagues need the data in Excel, filter it down with Power Query or split it. If you work with files like this every week, a little DuckDB or pandas pays off quickly. If the export is Parquet rather than CSV, inspect its schema and plot its numeric columns with the Parquet viewer. If it is a LabVIEW .tdms log, browse its groups and channels in the TDMS viewer.

Frequently asked questions

How do I open a CSV file with more than 1,048,576 rows?

Choose “Open and view” in the tool at the top of this page and pick the file, or drop it there. Your browser reads it on your own device and shows every row in a table, with each column’s type and statistics. In our tests a 2-million-row CSV opened in about 1 second, and a 10-million-row CSV opened and drew its first chart in about 3 seconds (measured in Chrome on a recent laptop, with a 423 MB CSV of 10,000,000 rows and 5 columns).

What is the maximum number of rows in Excel?

1,048,576 rows including the header, and 16,384 columns per worksheet, in every version since Excel 2007, including Microsoft 365. The older .xls format allows 65,536 rows.

Can I increase Excel’s row limit?

No. The limit is fixed in Excel’s grid and file format, and no setting or add-in changes it. The workaround inside Excel is to load the data into the Data Model with Power Query, which has no fixed row limit but does not show the rows in the grid.

Does Excel delete the rows it cannot show?

Not from the original file. Excel only loads the first 1,048,576 rows into the workbook. The extra rows are lost only if you save the truncated workbook over the original CSV, so always keep a copy of the export.

Can I get part of the file back into Excel?

Yes. Open the file in the tool at the top of this page, then sort and filter the table to the rows you need, or zoom a chart into a range. Under “Do more”, “Download the table as shown (CSV)” saves the table’s rows as you see them, sorted and filtered, with every column; “Download all columns (CSV)” saves every column and every row, the whole file; and “Download chart data (CSV)” saves the plotted columns for the rows in view. Excel opens the CSV whole as long as it has at most 1,048,576 rows, the header included.

Can Google Sheets open a CSV with more than a million rows?

Only if it has fewer than 20 columns. Google’s help page (checked 10 October 2026) says a spreadsheet holds “up to 20 million cells or 100MB”, counting every column: a CSV with 10 columns tops out at 2 million rows and one with 50 columns at 400,000, and a big file can reach 100 MB first.

How do I open a large CSV file on a Mac?

Apple Numbers and Excel for Mac both stop at around a million rows, and Excel’s Data Model is only fully supported on Windows. The command line (head, wc -l, split), DuckDB, Python and browser-based tools all work well on macOS.

Is it safe to use an online CSV viewer for confidential data?

Only if the file never leaves your computer. Some online tools send the file to their server and some read it in your browser. Tools that process data locally in the browser, such as MegaRows, read the file on your device instead, and you can confirm this in the Network tab of your browser’s developer tools.