JSON to CSV

Two formats with genuinely different shapes

The friction in this conversion is structural, not cosmetic. CSV is a grid: rows and columns, every row with the same fields, no hierarchy, no types — everything is ultimately text. JSON is a tree: objects can nest inside objects, arrays can hold arrays, different records can carry different fields, and values have real types. Flattening a tree into a grid necessarily loses information, and the interesting question is always which information you are prepared to lose. A flat array of uniform objects converts perfectly, because it was already a grid wearing JSON's clothes. Anything genuinely hierarchical requires a decision about how to project it, and that decision belongs to you rather than to a converter guessing.

The escaping rules that quietly destroy data

CSV's apparent simplicity hides a specification that a surprising amount of software gets wrong. A comma inside a value would otherwise look like a column boundary, so the value must be quoted. A double quote inside a quoted value must be escaped by doubling it. A line break inside a value is legal but only within quotes, and plenty of parsers fall over on it anyway. Get any of this wrong and you do not get an error — you get a spreadsheet where one row has silently split into two, or a column has shifted, and nobody notices until a report is wrong three weeks later. Correct quoting is the least glamorous and most important part of writing CSV.