CSV to JSON Converter

Why splitting on commas breaks

The CSV rules — set out in RFC 4180 — allow a field to be wrapped in double quotes, and a quoted field may contain the delimiter, a line break, or a quote character escaped by doubling it. All three appear constantly in real exports: "Sato, Yuki" is one field, not two, and a postal address routinely spans lines inside one cell.

A naive split produces extra columns, misaligned rows and silently wrong data, and it does so without any error message. That is the worst kind of bug: the output looks plausible. The parser here walks the text character by character and tracks whether it is inside quotes, which is the only way to get all four cases right.

The delimiter is not always a comma

In countries where the comma is the decimal separator — most of continental Europe, much of Latin America — Excel exports CSV with semicolons, because otherwise 1,5 would split into two fields. Tab-separated exports are common from databases and from Google Sheets. Pipes turn up in older enterprise systems.

Detection here counts candidate delimiters in the first line and takes the most frequent, which is right almost every time. When a header contains commas inside quoted text it can guess wrong, so the override is there.

Type conversion, and when to switch it off

With conversion on, 42 becomes the number 42, true becomes a boolean, and an empty cell becomes null. That is usually what you want, because JSON consumers expect real types rather than strings that look like them.

Switch it off when your data contains values that *look* numeric but are not. Postcodes, phone numbers, product codes and account numbers all lose leading zeros the moment they become numbers — 00123 turns into 123 and cannot be recovered. Long identifiers are worse: JavaScript numbers lose precision above 2^53, so a 19-digit ID silently changes its last few digits.

Going the other way

Converting JSON to CSV means flattening a tree into a grid, and some information cannot survive that. An array of objects becomes rows and columns cleanly. Nested objects and arrays are written as JSON text inside the cell, because there is no standard way to represent them otherwise.

Keys are collected across every row rather than taken from the first one, so an object with an extra field does not lose it. Any value containing a delimiter, a quote or a newline is quoted and its quotes doubled, which is what makes the output safe to reopen in a spreadsheet.

Opening the result in Excel

• Non-Latin characters look wrong. Excel assumes the local codepage unless the file starts with a UTF-8 byte-order mark. Import via Data → From Text rather than double-clicking the file.
• Everything lands in one column. Excel expected your locale's delimiter. Import and choose it explicitly.
• Leading zeros disappear. Excel is helping. Set the column type to Text during import.
• Long numbers become 1.23E+18. Same cause, same fix — import that column as Text.