CSV to JSON Converter

Because the field needs to be wrapped in double quotes, and whatever wrote the file either did not do that or whatever is reading it does not honour it. Under RFC 4180 a quoted field may contain the delimiter, line breaks and doubled quotes, all of which are data rather than structure. Rows with the wrong number of columns are reported here with their line numbers rather than being padded out, because a ragged row nearly always means a quoting problem upstream and hiding it turns a visible error into corrupted data.

Because something converted it to a number, and 01001 as a number is 1001. It is not recoverable afterwards, which is why type conversion here leaves anything with a leading zero as a string, along with values too long to hold in a floating point number exactly. Product codes, phone numbers and IDs all survive. You can also turn conversion off entirely and keep everything as strings.

It is detected, and you can override it. Comma is standard, but Excel in most of Europe writes semicolons because the comma is the decimal separator there, and exports from databases are often tab-separated. Detection works by finding which candidate gives the most consistent column count across the first few lines rather than by counting occurrences, since a comma-separated file whose data is full of semicolons would otherwise be misread.

Nested objects become dotted column names, so a user object with a name inside it becomes a user.name column. Arrays of simple values are joined into one cell rather than exploded into columns, because exploding gives a different column count on every row and produces a file no spreadsheet will open sensibly. The header is the union of every key across every record, not just the first one, so a field that appears only in later records is not silently dropped.

No. The parsing and conversion happen in your browser and no request carries the file. That matters here more than for most tools, because a CSV is usually a customer list, a transaction export or something else that has no business being pasted into a stranger's server.

A byte order mark. Excel writes one at the start of a UTF-8 CSV, and a parser that does not strip it makes it part of the first column's name, so a lookup for id fails against a key that is invisibly different. It is stripped here before anything else happens.