CSV Viewer & Editor

CSV is less simple than it looks

Comma-separated values sounds like a format that could be parsed with a single split, and that belief is the source of most CSV bugs in the wild. The complications are real: a field containing the delimiter must be quoted, a quoted field can contain line breaks, and a quote inside a quoted field is escaped by doubling it. RFC 4180 describes the common case, but it was written in 2005 to document existing practice rather than to define it, and plenty of software deviates. The practical consequence is that a parser has to be a character scanner tracking quote state, not a string split, and that a writer has to know when quoting is required.

Everything is text

A CSV file carries no type information whatsoever. The characters 01234 might be an identifier, a post code, or the number 1,234 written oddly — the file does not say, and cannot. Spreadsheet applications resolve that ambiguity by guessing on import, which is convenient until it is not: leading zeros vanish, long digit strings become floating-point approximations, and text that resembles a date is converted into one. The most infamous case is gene names in biology, where a substantial fraction of published spreadsheets contained corrupted identifiers because names like SEPT2 were silently turned into dates. This tool never guesses. Every field is the text it was, from open to download.

Editing without a spreadsheet

Reaching for a full spreadsheet application to fix three cells in a data export is a heavyweight solution with a specific risk: opening and re-saving a CSV in one of those applications applies its type guessing to every field in the file, not just the ones you touched, and its own quoting and encoding conventions on the way out. A file can come back materially different from how it went in, with changes nobody asked for and nobody reviewed. Editing the text directly — which is what this tool does — changes only what you change.