CSV to JSON Converter

Delimiter detection

"CSV" files use different separators depending on where they came from. Excel in much of Europe writes semicolons, database exports often use tabs, and some systems use pipes. With the delimiter on Auto, the converter tries comma, semicolon, tab and pipe and picks the one that splits every row into the same number of fields. The footer shows which one it chose; pick one yourself if a file is ambiguous, for example a single-column list.

Headers and keys

With First row is header on, each later row becomes an object whose keys are the header names. Blank headers become column_1, column_2 and so on, and repeated header names get a suffix (name, name_2) so no column overwrites another. Turn the option off to get an array of arrays instead, which suits files without a header or data you will process by position.

Rows with more or fewer cells than the header are kept: extra cells get column_N keys and missing cells are empty. A warning tells you how many rows were uneven, which usually points to an unquoted delimiter inside a value.

Type inference

CSV has no types; every cell is text. With Infer types on, cells that are written exactly like JSON numbers become numbers, true and false in any common capitalisation become booleans, and empty cells and null become null. Anything that would change if it were converted stays a string: values with leading zeros like ZIP codes, phone numbers and account numbers, and whole numbers too long to store exactly, such as 19-digit IDs. Dates stay strings too, because JSON has no date type. Turn inference off to keep every value as the exact text from the file.

Quoted cells

Cells are read following RFC 4180. A cell wrapped in double quotes can contain the delimiter and even line breaks, and two double quotes in a row stand for one quote character. If a quote is opened and never closed, the converter stops with an error that names the row, instead of quietly merging the rest of the file into one cell.

Nested output

Tick Nest dot paths to rebuild structure from flattened headers: address.city and address.zip become an address object, and tags.0, tags.1 become an array. That makes this the reverse of JSON to CSV. A byte order mark at the start of the file, which Excel adds to UTF-8 exports, is removed automatically.