An Excel workbook is a rich container. A single file can hold several worksheets, formulas that recalculate as values change, cell formatting and colors, number and date types, and even charts. A CSV file is the opposite: one plain-text table where each line is a row and commas separate the fields. Converting Excel to CSV is therefore a flattening, not a simple copy, and knowing what survives that flattening saves you from surprises after the file lands in another program.
You reach for CSV when a downstream system wants raw tabular data and nothing else: a database import wizard, a Python or R script, an analytics tool, or another team that may not use a spreadsheet app at all. The CSV format is the universal exchange currency for tables precisely because it carries only the values, which is also why the conversion drops everything else. The details that decide whether that plain-text file imports cleanly, encoding, delimiters, and quoting, are where this guide spends most of its attention.
What changes when Excel becomes CSV
A spreadsheet stores machinery around your data. Formulas compute results, formatting controls how numbers appear, cell types distinguish a date from the text that looks like one, and multiple sheets live side by side in one file. CSV can represent none of that. When you export, each formula is replaced by its last computed value, formatting is stripped away, and only a single sheet is written out, so you choose which one. Merged cells are flattened and charts are simply gone.
That reduction is the point. If you needed formulas and formatting preserved, you would keep the workbook. You convert to CSV specifically because the receiving tool wants a clean, predictable grid of values it can read without any spreadsheet-specific handling.
Two formats: a spreadsheet engine and a plain grid
The modern Excel workbook (.xlsx) is an Office Open XML file: a ZIP archive containing XML documents that describe each worksheet, a shared string table, style definitions, and a calculation chain that records how formulas depend on one another. The older .xls is a binary format from the same lineage. In both cases the file is far more than a table of values, it is a small computational document, storing not just what each cell shows but how it was computed, how it is formatted, and how the sheets relate.
CSV, by contrast, has essentially no structure beyond rows and fields. It is comma-separated values: lines of text, each line a record, fields within a record separated by a delimiter. There is no notion of a formula, a data type, a color, or a second sheet. That radical simplicity is exactly why almost every data tool in existence can read a CSV, and exactly why so much has to be discarded to produce one.
What the CSV standard actually says
For decades CSV was a folk format with no single owner, which is the root of many import headaches. It was eventually documented in IETF RFC 4180, which defines a common form: records separated by a carriage-return/line-feed pair, fields separated by commas, an optional header line, and rules for when and how to quote fields.[1] The Library of Congress catalogs this RFC 4180 form as the reference definition of the format.[2]
The important caveat is that RFC 4180 describes common practice rather than mandating it, and it is explicitly informational.[1] Real-world CSV files vary: some use a semicolon or tab instead of a comma, some use a single newline rather than the RFC's carriage-return/line-feed, and character encoding is not part of the core specification at all. Knowing that the standard exists but is loosely followed is what lets you anticipate the two problems that actually break imports, encoding and delimiters, which the next sections cover.
What is kept and what is lost
Before you export, it helps to see exactly which parts of the workbook make the trip and which do not.
| In your Excel file | In the CSV |
|---|---|
| Cell values (text, numbers, dates) | Kept as plain text |
| Formulas | Replaced by their last calculated value |
| Multiple worksheets | Only one sheet per CSV; export each separately |
| Formatting, colors, fonts | Lost |
| Charts, images, pivot tables | Lost |
| Merged cells | Flattened; the value lands in the top-left cell |
The single-sheet limitation deserves a moment of planning rather than a shrug. A workbook that models something with several related tables, say a data sheet, a lookup sheet, and a summary, cannot travel as one CSV, because a CSV is one table by definition. You export each sheet to its own file and reassemble the relationship in the destination tool, whether that is a set of related database tables or several data frames in a script. Recognizing this before you export avoids the common mistake of exporting only the visible sheet and later discovering the supporting data never came along.
Encoding and delimiter pitfalls
Most "my CSV looks broken" complaints trace back to a handful of settings rather than the data itself. Getting these right up front produces a file that imports cleanly on the first try.
The first is character encoding. CSV is plain text, but plain text still has to be encoded, and RFC 4180 does not fix a single encoding.[1] If a file written as UTF-8 is opened by a tool that assumes an older single-byte code page, accented letters, curly quotes, and non-Latin scripts turn into garbled sequences. The reliable choice is UTF-8, which can represent any character; the wrinkle is that some Windows tools expect a byte-order mark at the start of the file to recognize UTF-8, so if characters look wrong, an added or removed BOM is a common fix.
The second is the delimiter. The standard uses a comma, but locales that use the comma as a decimal separator (much of Europe) commonly export CSV with a semicolon instead, and some data uses tabs. If every row of your imported file lands in a single column, the receiving tool is splitting on the wrong delimiter. Match the delimiter to what the target expects rather than assuming a comma.
Dates are often serialized to a text form, values containing commas must be wrapped in quotes, and codes like ZIP codes and long ID numbers can lose leading zeros or shift into scientific notation because CSV records only the value, not the cell's number format. Choose UTF-8 encoding so accented and non-Latin characters survive, and confirm the delimiter your target tool expects.
Quoting, commas, and line breaks
The one genuinely subtle part of the CSV format is how it handles field values that themselves contain the delimiter. RFC 4180 defines the rule: any field that contains a comma, a double quote, or a line break must be enclosed in double quotes, and a literal double quote inside such a field is escaped by doubling it.[1] So the value Smith, John is written "Smith, John", and a value like 6" pipe becomes "6"" pipe". This quoting is what lets a field safely contain the very character that separates fields, and it is why a correctly written CSV never loses a column when a cell happens to contain a comma.
A correct converter applies these quoting rules automatically, which is exactly the kind of edge case a hand-rolled export often gets wrong. If you have ever seen a name with a comma spill into the next column, you have seen quoting done incorrectly. It is worth knowing the rule exists so you can recognize a broken file when the fault is the writer, not your data.
Matching the CSV to the tool that reads it
The settings that matter change depending on what will consume the file, so it helps to think about the destination before exporting.
| Destination | What to get right |
|---|---|
| Database import wizard | UTF-8 encoding, a header row matching column names, and a delimiter the importer expects; watch for text codes losing leading zeros. |
| Python (pandas) or R script | UTF-8 and correct quoting; these tools follow RFC 4180 quoting closely and will misparse a file that quotes badly. |
| A European colleague's spreadsheet | The delimiter may need to be a semicolon, since a comma is the decimal separator in many locales. |
| A web form or API bulk upload | UTF-8 without surprises, and confirm whether the service wants a header row at all. |
| Plain text inspection | Any consistent delimiter works; just keep it uniform across every row. |
The common thread is that a CSV is only as good as the assumptions of the tool reading it. Because the format itself is loosely standardized, the safest habit is to know your reader's expectations, encoding, delimiter, and whether it wants a header, and produce the file to match, rather than exporting once and hoping.
Convert Excel to CSV step by step
Open the converter
Open the FileFormer document converter. It runs entirely in your browser, so you can begin without installing anything or creating an account.
Add your Excel file
Drag your .xlsx or .xls workbook into the drop area or browse to select it. The file is read locally by your browser rather than sent anywhere.
Choose the sheet and encoding
Pick the single worksheet you want as CSV, since only one sheet exports at a time, and confirm UTF-8 encoding so accented characters and symbols stay intact.
Export and download
Convert and download the CSV. Because the work happens on your device, the workbook, which may hold customer or financial data, is never uploaded to a server.
Convert Excel to CSV in your browser
The document converter turns .xlsx and .xls into clean, UTF-8 CSV with correct quoting, entirely on-device with nothing uploaded.
Key takeaways
- CSV keeps only cell values from one sheet; formulas become their last computed value and formatting, charts, and extra sheets are dropped.
- Export each worksheet separately, since a CSV holds a single table.
- Save as UTF-8 and match the delimiter your target tool expects to avoid garbled characters and misaligned columns.
- The conversion runs in your browser with no upload and no sign-up.