CSV viewer

Drop the file in and it opens. There's no row limit imposed here, and nothing is uploaded, so the constraint is your own machine rather than a server or a spreadsheet's maximum.

Excel's own limit is 1,048,576 rows, and files well below that become slow long before they become impossible.

What Excel does to a CSV that this doesn't

This is the more important reason to look at a CSV somewhere else first. Excel interprets values as it opens them, and several of those interpretations are destructive.

Long numbers become scientific notation. A 16-digit account reference displays as 1.23457E+15, and if the file is saved from that state, the digits are genuinely gone.

Leading zeros disappear. A postcode or product code of 01234 becomes 1234.

Things that look like dates become dates. The classic case is gene names in biology, where SEPT2 becomes a September date, which was widespread enough that the naming committee renamed the genes. Version numbers and part codes hit the same problem.

Long identifiers lose precision. Anything past 15 significant digits is rounded, silently.

None of that happens here, because nothing is interpreted. The file is shown as it is.

That makes this a good first stop when you've been sent a CSV and need to know what's actually in it before a spreadsheet has an opinion about it.

When the columns come out wrong

Almost always the separator.

The C in CSV means comma, and plenty of files use something else. Semicolons are standard across much of Europe, because those locales use a comma as the decimal point and a file can't use it for both. Tabs are common in exports from databases.

If your data appears in one long column, the separator is wrong. If it's split in odd places, the file may be using a separator that also appears inside the values.

The separator is detected from the file here and shown in the header, so you can see what it decided.

Quotes and the fields that contain commas

A value containing the separator has to be quoted: "Smith, John" is one field, not two. A quote inside a quoted field is doubled: "She said ""hello""".

Files that don't follow that convention are broken, and every reader will disagree about what they mean. If a row has more fields than the header, an unquoted separator inside a value is the usual cause.

Encoding, and the odd characters at the start

If the first heading begins with , the file has a byte order mark and whatever opened it read it as the wrong encoding. If accented characters look like é instead of é, a UTF-8 file has been read as Latin-1.

Both are display problems rather than damage, and both are fixed by opening the file as UTF-8. Excel is the usual culprit on Windows, and its import dialogue lets you set the encoding explicitly.

Very large files

The whole file is read into memory, so a very large one can be slow or can fail where a smaller one succeeds. Several hundred megabytes is where a browser tab starts to struggle.

For files beyond that, command line tools are the honest answer. head, wc -l and csvkit will tell you what you need to know without loading everything.

Turning it into something else

Converting to JSON keeps numbers as numbers and turns empty cells into null, which is what most code expects.

For anything going back into a spreadsheet, keep it as CSV and import rather than opening. Excel's import path lets you mark a column as text, which prevents every one of the problems above.

Nothing is uploaded. The file is read in your browser, which matters here because CSVs tend to hold exports of exactly the data nobody wants on a server.