Open a CSV file, see what is in every column, then search and sort it.
A CSV file is a table stored as plain text, one row per line, with the values separated by commas. Open one in a text editor and you get a wall of characters. Open it in Excel and you might get a dialog about separators, a column of order numbers turned into scientific notation, and every leading zero quietly stripped off. A CSV viewer sits in between: it reads the file as a table and shows it to you without changing anything.
We rebuilt this one in 2026 around a question the other viewers do not answer, which is what is actually in the file. Every column header here carries its detected type, how much of the column is filled in, how many distinct values it holds, and the range if it is numeric. That is usually the first thing you want to know about a file somebody just handed you, and it takes about a second to read. The file itself never leaves your machine, because the parsing all happens in JavaScript in your browser.
>100 or <=50, and (blank) finds empty cells anywhere.A CSV turns up at the end of almost every export button ever written, so the reasons for opening one vary a lot.
Types come from the values themselves, sampled across the first 5,000 rows. A column is only called a number if every value in it is one, so a single "N/A" in a price column shows up as text and mixed, which is exactly the warning you want.
| Type | Looks like | What it changes |
|---|---|---|
| Integer | 42, -7, 1,024 | Sorts numerically, shows min, max and average |
| Number | 19.99, 1.5e3 | Sorts numerically, right aligned in a monospace font |
| Currency | $19.99, £5, €12,50 | Sorts by the number inside the symbol |
| Percent | 42%, -3.5% | Sorts by the number, keeps the sign |
| Date | 2026-08-05, 05/08/2026, ISO 8601 | Sorts chronologically rather than alphabetically |
| Time | 09:30, 17:45:00 | Sorts chronologically |
| Boolean | true, no, Y | Flagged so you know the column is a flag |
ada@example.com | Flagged, useful for spotting a column of addresses | |
| URL | https://example.com/x | Flagged and colored |
| UUID | 8f14e45f-ceea-... | Flagged as an identifier |
| IP address | 192.168.1.10 | Flagged as an identifier |
| JSON | {"a":1} | Flagged, common in event exports |
| Text | Anything else | Sorts alphabetically, case insensitively |
Drop it on the box at the top of this page. The file is read by your browser and shown as a sortable table straight away, with no install and no account. This also avoids the things Excel does to a CSV on the way in, like turning 00123 into 123, reading a long order number as scientific notation, and converting anything that looks vaguely like a date. What you see here is what the file actually says.
No. The whole thing runs in JavaScript in your browser, so the file is read off your own disk and never crosses the network. You can prove it: open your browser's network tab, load a file, and watch for a request that never comes. Turning off your internet connection after the page has loaded also works, and the tool keeps going.
The practical ceiling is your computer's memory, because the file is held in the browser rather than streamed from a server. Files in the tens of megabytes open in a second or two on an ordinary laptop, and hundreds of thousands of rows are fine because only the rows on screen are ever drawn. We stop at 300 MB and say so rather than freezing the tab. If your file is bigger than that, split it first or filter it down with a command line tool.
Because the separator is not a comma. Files exported in Europe usually use semicolons, since the comma is the decimal point there, and files out of a database often use tabs. The separator is detected automatically here by looking for the character that gives every row the same number of fields, but you can pick one by hand from the dropdown if the file is unusual.
Yes, and the sorting knows what each column holds. Click a header to sort, click again to reverse, click a third time to go back to the file's own order. Numbers sort numerically, so 9 comes before 10 rather than after it, and dates sort chronologically rather than as strings. Empty cells always sink to the bottom whichever direction you pick. For filtering, the box at the top searches every column and the small box in each header filters that column alone.
No. Every value is kept as the text it was in the file, so 00123 stays 00123 and a phone number starting with a zero survives. The tool even points out which columns have leading zeros, because those are the ones that would be damaged if you opened the same file in a spreadsheet. If you need to convert the file rather than read it, our CSV to JSON converter keeps them too.
It is the fill rate, the share of rows where that column has a value. A column at 12% is mostly empty, which usually means it is optional, badly populated, or the wrong column entirely. It is the fastest way to find the fields nobody actually filled in, and it is measured across the first 5,000 rows.
Yes. Tab, semicolon, pipe, colon and tilde are all recognized, and any of them can be chosen by hand. The file extension does not matter, so a .txt file holding a table opens the same way a .csv does.
They are shown rather than dropped. Short rows are padded with empty cells, long rows are trimmed to the width of the table, and a note under the table tells you how many rows were affected and which line the first one is on. This is nearly always caused by an unescaped comma or a stray quote somewhere upstream, and knowing the line number is what lets you go and fix it.
Yes, and what you export is what you are looking at, filters and hidden columns included. Download it as CSV, as JSON with numbers and booleans converted to real types, or as a set of SQL INSERT statements. There is also a copy button that gives you a Markdown table, which is handy for a pull request or a ticket. For a proper conversion with more control, try our CSV to SQL converter.