CSV Validator Online

Paste or drop a CSV and get a list of what is wrong with it: rows with the wrong number of fields, duplicate or blank header names, columns whose values disagree about their type, and characters that suggest an encoding problem.

Input

Read on this page. Nothing is uploaded.

Findings

Paste or drop a CSV. It is checked on this page and never uploaded.

What is checked

Row width against the header. A row with more fields than the header usually means an unquoted delimiter inside a value; a row with fewer usually means a truncated line.

Header names. Blank headers and duplicate headers both cause silent data loss downstream, because most tools key columns by name.

Type consistency per column. A column that is numeric in 998 rows and text in two is worth looking at, and the two are usually where the export went wrong.

Replacement characters. A U+FFFD in the text means the file was decoded with the wrong encoding and those characters are already lost.

There is no CSV schema

CSV has no way to declare what a valid file looks like, so validation is a set of heuristics rather than a check against a specification. Everything reported here is something that is usually a mistake, not something that is definitionally illegal.

Questions

My file is reported as ragged but looks fine.

A value containing an unquoted delimiter is the usual cause. It looks fine in a text editor because the extra comma is invisible among the others, and it breaks the row because the parser has no way to know it is data.

Other tools