Convert nested JSON to CSV without hidden data loss

A table needs a row boundary

The converter accepts an array of objects, with each object becoming one row. If an API response wraps records inside a results property, extract that array first. A single object can be wrapped in square brackets. An empty array provides no fields to use as a table schema.

Follow the flattening rule

[{"id":"00123","address":{"city":"Lahore"},"tags":["writer","editor"]}]

With flattening enabled, address.city becomes a column. The tags array stays JSON text in one cell. It does not create two rows. With flattening disabled, the address object also stays JSON text.

If the same record contains both a literal key address.city and a nested address object with city, flattening would create a collision. The converter blocks that case so one value cannot overwrite the other.

Some distinctions cannot survive

Missing fields and explicit null both become empty cells. A CSV also does not carry a rich type schema. Consequently, converting back to JSON cannot reliably reconstruct the original structure or null/missing distinction. Arrays stay serialized, but a downstream spreadsheet may still alter their quoting or values on save.

JSON number tokens are kept as text during conversion, so 9007199254740993 is not rounded by the browser. The receiving application must also preserve it; import identifier columns as text.

Choose and inspect columns

Fields are collected across all records in first-seen order. Use optional one-based column positions to select or reorder the output, then inspect the header and a few rows. CSV quoting protects embedded separators and newlines. Formula-like cells are prefixed in spreadsheet-safe mode.

Validate JSON first when a syntax error prevents conversion. A valid syntax result still says nothing about whether the chosen array contains the business records you intended.