What the converter expects
JSON→CSV works best on the shape CSV was built for: an array of objects with consistent keys. The first object’s keys become the header row; each object becomes a data row. If your API wraps the data ({ "results": [ ... ] }), extract the array first — data.map is not a function is the classic error from handing the converter a wrapping object instead of the array inside it.
[ { "id": 1, "name": "Alice" }, { "id": 2, "name": "Bob" } ]becomes a clean two-column, two-row CSV.
The comma-and-quote problem, handled
CSV’s fatal weakness is that its delimiter (the comma) frequently appears inside values — addresses, descriptions, lists. The RFC 4180 standard solves this with quoting, and the converter applies it automatically:
- A value containing a comma, quote, or newline is wrapped in double quotes.
- A literal double quote inside a value is escaped by doubling it (
"→"").
So Smith, John becomes "Smith, John" and He said "hi" becomes "He said ""hi""". This is why you should never split CSV on commas with a naive split(',') — a compliant parser respects the quotes.
The UTF-8 / Excel encoding trap
Open a CSV with accented names or non-Latin text in Excel and you may see é where é should be. The file is correct UTF-8; Excel (especially on Windows) assumes a legacy locale encoding instead. Two reliable fixes:
| Fix | How |
|---|---|
| Import explicitly | Data → From Text/CSV → choose UTF-8 |
| Add a BOM | A UTF-8 byte-order mark tells Excel the encoding |
Google Sheets and modern tools handle UTF-8 correctly without this dance — the problem is specific to Excel’s import heuristics.
Going back: CSV to JSON and the type question
The reverse direction has one inherent limitation: CSV has no type system, so every value arrives as a string. "42", "true", and "2026-01-01" all come back as strings unless you post-process them. If your downstream code needs real numbers and booleans, add a typing pass after conversion — or, better, validate against a known schema so each field is coerced to its intended type deliberately rather than guessed.