CSV Guide

CSV delimiters: comma, semicolon, tab, and pipe

The separator is part of the file format. Changing it safely requires parsing fields first, not blind search-and-replace.

A CSV file does not always use commas

The name “comma-separated values” suggests one separator, but real-world exports often use semicolons, tabs, or pipes. Tab-separated files are usually called TSV, yet many systems still use a .csv extension for other delimiters.

Why semicolons are common in spreadsheet exports

Some regional settings use a comma as the decimal separator. A semicolon can then be used as the field separator so a value such as 19,95 does not conflict with column boundaries. This is one reason a file that looks fine in a local spreadsheet can import as a single column elsewhere.

Delimiter detection is based on consistent structure

A detector looks at several candidate characters and asks which one produces a stable number of fields across multiple rows. One row is not enough evidence. Free-text values, addresses, and descriptions can contain commas that are not separators.

Quoted delimiters do not split fields

sku,title,description
A1,"Blue, Large","Cotton shirt"

The comma inside "Blue, Large" belongs to the field because the value is quoted. A correct delimiter conversion must parse the CSV first and then serialize it again. Replacing every semicolon or comma in the raw text can corrupt valid field contents.

Safe delimiter conversion

  1. Detect or confirm the source delimiter.
  2. Parse the CSV according to quoting rules.
  3. Verify that rows resolve to a consistent field count.
  4. Serialize the parsed fields using the destination delimiter.
  5. Quote fields when the destination delimiter appears inside a value.

Signs that the delimiter is wrong

  • The importer shows one giant column instead of many.
  • The header line appears as one field.
  • Every row has the same text but the parser reports one column.
  • A semicolon file opens correctly in a regional spreadsheet but not in a web importer.

When the delimiter is not the only problem

If rows resolve to different column counts after the correct delimiter is selected, the file likely has a quoting or structural issue too. Use the Column Count Checker and Broken Quote Checker.

Try it: Detect and normalize a CSV delimiter locally in your browser.