CSV to Excel

The two things Excel does to your data on import

Opening a CSV directly in Excel is quick and it is also where a great deal of data quietly dies. Two conversions are responsible for most of it.

The first is leading zeros. A column of order references like 00123, product codes like 007, or Indian PIN codes and postcodes beginning with a zero are read as numbers, and the zeros are discarded. 00123 becomes 123, and there is no undo that brings them back — the information is simply gone from the sheet.

The second is long digit strings. A 16-digit card number, a 19-digit transaction reference, a long IMEI or account number exceeds the 15 significant digits a spreadsheet number can hold, so Excel stores an approximation and displays it in scientific notation. 4532156789012345 becomes 4.53216E+15, and the last digit is lost permanently even after you reformat the cell.

This converter writes the workbook itself and applies a deliberately strict rule about what becomes a number: a plain optional-minus integer or decimal, no leading zeros, no more than 15 significant digits. Anything else — leading zeros, thousands separators, currency symbols, percentages, dates, long references — is written as text and survives exactly as it appears in your file. Genuine numbers still arrive as numbers, so SUM and AVERAGE work on the columns where they should.

Parsing the CSV correctly in the first place

Before any of that can matter, the file has to be read properly. The rules are RFC 4180: a field wrapped in double quotes may contain the delimiter, may contain newlines, and represents a literal double quote as two of them. A converter that splits each line on commas mangles all three, and it does so silently — the row count looks plausible and the columns are wrong.

The parser here is a state machine that tracks whether it is inside quotes, so `"Smith, John"` stays one field, a quoted address spanning two lines stays one cell, and `"He said ""yes"""` comes through as `He said "yes"`.

The delimiter is detected from the header line — comma, semicolon or tab — and can be overridden if the guess is wrong, which happens with files that have a comma in the first heading. The preview above the button shows the first rows as they will land in the sheet, so a misdetected delimiter is obvious before you download anything rather than after you have imported it somewhere.

What the tool writes

A genuine Office Open XML workbook: a content-type map, package relationships, a workbook part, a worksheet and a style table. It opens in Excel, LibreOffice, Numbers and Google Sheets as a native file, with a proper sheet name taken from your filename — trimmed to the 31 characters Excel allows and stripped of the characters it forbids in a sheet name.

That is worth distinguishing from what several "converters" do, which is rename the file or wrap it in a single-sheet HTML table that Excel happens to open. Both produce a file that works until something else has to read it, at which point it turns out not to be a spreadsheet at all.

Very large files are the practical limit rather than any imposed one. A CSV of a few hundred thousand rows will build, but it takes time and memory, and a spreadsheet is rarely the right destination for data at that scale in any case.

Converting without uploading

CSV exports are what systems produce when you ask them for your data: a bank's transaction history, a payroll run, a CRM's contact list, an analytics export. They are almost by definition the files you would least like to leave on a stranger's server, and yet uploading them to a conversion site is the normal way this job gets done.

This runs in your browser. The CSV is read from your disk by the file picker, parsed in memory, and the workbook is assembled and handed straight back as a download. There is no server, no upload progress bar and no retention policy to read, and the whole thing works with the network disconnected once the page has loaded.

If what you actually wanted was the reverse trip, the Excel to CSV tool on this site handles it with the same care about quoting and dates — and the CSV to JSON tool covers the case where the destination is an API rather than a spreadsheet.