How to convert JSON to CSV on iPhone

Convexy converts JSON that is an array of objects at the top level. Like this:

[
  { "name": "Ada",   "city": "London",   "age": 36 },
  { "name": "Grace", "city": "New York", "age": 45 }
]

which becomes:

age,city,name
36,London,Ada
45,New York,Grace

Each object is a row. The columns are the union of every key found across all the objects, sorted alphabetically — so if some records have a key that others lack, the column still appears, and the records missing it get an empty cell. Nothing is dropped for being inconsistent.

Values that contain a comma, a quote or a newline are wrapped in quotes and their internal quotes doubled, which is the correct CSV escaping. The output opens cleanly in Excel, Numbers and Google Sheets.

A CSV cell holds one value. JSON values can be objects and arrays, which hold more values, which can hold more values. There is no honest way to put a tree into a cell.

Real-world JSON is full of this:

[
  {
    "name": "Ada",
    "address": { "city": "London", "zip": "N1" },
    "tags": ["admin", "beta"]
  }
]

Every tool has to pick a lie to tell here. The common ones are flattening (invent columns called address.city and address.zip, and give up on tags having any number of entries) or dropping the nested fields entirely.

What Convexy does: the record still converts, and the nested value is written into its single cell as a text dump of the structure. You will not lose the row, and you will not silently lose the field — but what lands in that cell is a representation of the object, not a clean value, and it is not something you would want to hand to a spreadsheet user. Treat a nested field as a signal that the file needs reshaping before conversion, not after.

The practical fix, if you control the JSON: flatten it at the source into one level of keys and simple values. That is the shape CSV was designed for, and it is the only shape that converts without a compromise.

The most common reason this conversion fails is that the data you want is not at the top of the file. API responses almost always wrap it:

{
  "status": "ok",
  "page": 1,
  "results": [ { ... }, { ... } ]
}

The top level here is an object, not an array, so Convexy rejects it — the array you actually want is buried under results. Convexy will not go looking for it, and will not guess which key holds your records. Extract the array yourself first — pull out the value of results into its own file so the file begins with [ — and then convert. It is a two-second edit in any text editor, and it means you decided which data was the data, rather than a converter guessing.

• Columns come out alphabetical, not in the order the keys appear in the JSON. JSON objects have no guaranteed key order, so there is no original order to preserve. Reorder them in the spreadsheet if it matters.
• Missing keys become empty cells, so every row has the same number of columns and the file is a valid grid.
• The file is written as UTF-8. If Excel on Windows shows mojibake where the accents should be, use Excel's Import Data flow and choose UTF-8 rather than double-clicking the file — that is Excel's fault, not the file's.

CSV to JSON is supported and has its own set of honest caveats — chiefly that every value comes out as a string, because CSV has no types to preserve. Note that a JSON to CSV to JSON round trip will not give you back your original file: the types and any nesting are gone after the first hop, and CSV cannot carry them.