JSON to CSV converter

CSV is a flat grid, and JSON is a tree, so a converter has to make a few choices. These are the ones this tool makes, and the options that change them:

Values are written the way RFC 4180 describes CSV: a field that contains the delimiter, a double quote or a line break is wrapped in double quotes, and a quote inside it is written twice ("Chen ""CJ"" Jun"). Lines end with CRLF by default, as the RFC specifies; switch it off for LF if a Unix tool will read the file.

• Accents and emoji look broken. Excel guesses the encoding of a CSV file. The download starts with a UTF-8 byte order mark (BOM), which tells Excel the file is UTF-8. Google Sheets reads UTF-8 either way.
• Everything lands in one column. Excel splits on the list separator of your regional settings, which is a semicolon in many European locales. Pick Semicolon, or import the file with Data, From Text/CSV and choose the delimiter there.
• Leading zeros disappear. Excel turns 02134 into the number 2134 when it opens a CSV. The CSV still holds the zeros; import the column as Text to keep them.
• Data from users can run as a formula. A cell that starts with =, +, - or @ may be treated as a formula, which OWASP calls CSV injection. The "Protect against formula injection" option prefixes such cells with a single quote, as OWASP suggests. OWASP also notes that Excel can drop that protection after the file is saved and opened again, so treat exported user data with care.

JavaScript, no library, for an array of flat objects. Nested values are written as JSON text:

// rows: an array of flat objects
const cols = [...new Set(rows.flatMap(Object.keys))];
const cell = v => {
  const s = v == null ? '' : typeof v === 'object' ? JSON.stringify(v) : String(v);
  return /[",\r\n]/.test(s) ? '"' + s.replace(/"/g, '""') + '"' : s;
};
const csv = [cols, ...rows.map(r => cols.map(c => r[c]))]
  .map(r => r.map(cell).join(','))
  .join('\r\n');

Python with the standard library. Open the file with newline="", as the csv module docs say, or you get blank lines between rows on Windows:

import csv, json

with open("users.json", encoding="utf-8") as f:
    rows = json.load(f)

cols = list(dict.fromkeys(k for r in rows for k in r))
with open("users.csv", "w", newline="", encoding="utf-8") as f:
    w = csv.DictWriter(f, fieldnames=cols)
    w.writeheader()
    w.writerows(rows)

pandas flattens nested objects with json_normalize, using a dot as the separator by default:

import json
import pandas as pd

data = json.load(open("users.json", encoding="utf-8"))
df = pd.json_normalize(data)          # {"address": {"city": ..}} -> address.city
df.to_csv("users.csv", index=False)

jq on the command line. @csv takes an array and writes it as one CSV line, with strings in double quotes:

# One row per object, with a header row; @csv quotes strings and doubles quotes
jq -r '(.[0] | keys_unsorted) as $k | $k, (.[] | [.[$k[]]]) | @csv' users.json > users.csv

• Invalid JSON. Single quotes, trailing commas, comments and unquoted keys are not JSON. The error shows the line and column. To fix and pretty-print the JSON first, use the JSON formatter.
• Very large integers. JSON itself has no size limit, but RFC 8259 notes that numbers are only portable within the range of IEEE 754 doubles. JavaScript, and this tool, read integers above 9,007,199,254,740,991 approximately. Keep long IDs as strings in the JSON.
• Deeply nested arrays of objects. An order with ten line items is really two tables. Paste the inner array on its own to get one row per item, or use Column per item to keep one row per order.

Going the other way? Use CSV to JSON, or open a CSV file as a table with the CSV viewer. Looking for data to convert? Browse free public APIs.