Nested JSON ⇄ CSV (Flatten)

Worked example: a product catalogue

[
  { "sku": "K-102", "price": { "mrp": 1299, "sale": 999 },
    "sizes": ["S", "M", "L"], "stock": { "2025": 40, "2026": 12 } },
  { "sku": "K-205", "price": { "mrp": 899, "sale": 899 },
    "sizes": ["M"], "stock": { "2026": 7 } }
]

sku,price.mrp,price.sale,sizes.0,sizes.1,sizes.2,stock.2025,stock.2026
K-102,1299,999,S,M,L,40,12
K-205,899,899,M,,,,7

Converting back rebuilds K-102 exactly. K-205 gets its single size back as ["M"], because empty cells under numbered columns are skipped, but its stock comes back as {"2025": "", "2026": 7}: an empty cell under an ordinary key is kept as an empty string. The year keys stay object keys. Numbered segments only become an array when they start at 0, as sizes.0 does, so stock.2025 doesn't turn into an array with two thousand empty slots.

Editing the sheet before converting back

• New field: add a column such as price.gst and it appears inside price on every record, as "" where you left it blank.
• New array item: add sizes.3 after sizes.2. Filling it for K-102 gives four sizes.
• Removing an item: clearing sizes.1 while sizes.2 is still filled leaves a gap, which comes back as ["S", null, "L"]. Move the later values left instead.
• Renaming a header changes the structure: price.sale renamed to offer.price moves the value into a new offer object.

Frequently Asked Questions