CSV converter

CSV means comma-separated values, but real CSV is a family of delimited text files rather than one perfectly consistent format. RFC 4180 describes common comma, quote, and CRLF rules, yet everyday files use semicolons, tabs, pipes, regional decimal conventions, bare quotes, embedded line breaks, comments, different encodings, and inconsistent row lengths. CSV is popular because it is easy to inspect, stream, version, email, and import into spreadsheets or databases. It carries only text fields in rows. It does not carry formulas, cell types, styles, multiple sheets, charts, constraints, or schema unless another system supplies them.

Variants and dialect choices matter more than the extension. A European export might use semicolons because commas are decimal separators. A tab-separated file may be named .csv by a vendor. Excel may save CSV in a local code page, UTF-8 with BOM, or UTF-16 depending on version and platform. Quoted fields can contain commas, quotes, and newlines, but many naive parsers split lines incorrectly. Header rows are conventions, not guarantees. Some systems treat blank fields as null, others as empty strings. Numeric-looking identifiers can lose leading zeroes if opened in a spreadsheet without import settings.

CSV is read and written by Excel, Google Sheets, databases, pandas, R, business intelligence tools, accounting systems, CRMs, ERPs, ad platforms, analytics exports, and API ingestion jobs. Those readers infer types differently. One tool sees 1-2 as a date, another sees text. One tool accepts unescaped newlines, another rejects the file. A database import may require exact column counts and encodings, while a spreadsheet silently repairs rows. Reliable conversion records the dialect: delimiter, quote character, newline, encoding, header handling, and target schema.

Conversion gotchas are common but avoidable. CSV to XLSX can preserve a table and prevent accidental delimiter misreads, but it cannot recover formulas or styles that never existed. CSV to JSON must decide whether rows become objects, arrays, typed values, or strings. CSV to SQL needs safe table and column names, escaping, and type decisions. CSV to Markdown works only for reasonably small tables. Before converting, inspect sample rows with quotes, commas, newlines, non-ASCII characters, blank values, duplicate headers, and very large numbers. For production pipelines, reject ragged rows early and keep a schema contract next to the file.

Build CSV processes like contracts, not casual exports. Record who produces the file, which delimiter and encoding are expected, whether a header is mandatory, which columns are required, and how nulls are represented. Validate row counts and column counts before conversion. Compare totals or checksums after conversion when money, inventory, or compliance data is involved. Avoid opening production CSVs directly in spreadsheets unless import settings are controlled, because Excel may rewrite dates, identifiers, and large numbers on save. When exchanging files with partners, include a small sample and a data dictionary so the delimiter and type assumptions are not rediscovered during every import or onboarding cycle. Version that contract when columns change and reject unknown extras intentionally, especially during every scheduled vendor import run and production validation job checks.

Official CSV specification