CSV to JSON converter

How values are typed

CSV has no types: every cell is text. With “Detect numbers, booleans & nulls” on, 42 and 3.14 become JSON numbers, true/false become booleans, and empty cells or the word null become null.

Codes with a leading zero such as ZIP 02134 or product 007 stay strings, because turning them into numbers would silently drop the zero. Integers larger than JavaScript’s safe limit (9,007,199,254,740,991) also stay strings so IDs are not rounded.

Nested JSON from flat columns

Turn on dot-header nesting and a column named address.city becomes {"address": {"city": …}}. A header like tags.0 creates an array. This is the exact reverse of the flattening the JSON to CSV converter does, so you can round-trip data.

Which output shape to choose

Array of objects is what most APIs and JavaScript code expect. Array of arrays keeps it compact and preserves column order. JSON Lines (one object per line, also called NDJSON) streams well into BigQuery, Elasticsearch bulk loads and log pipelines. Keyed by first column gives you a lookup table such as {"1": {...}, "2": {...}}.

Convert CSV to JSON in code

For repeated or automated jobs, a few lines of code do the same thing. Python’s standard library csv.DictReader reads each row as a dictionary keyed by the header; json.dump writes it out. Note that DictReader leaves every value as a string, unlike the type detection above.

import csv, json

with open('input.csv', newline='', encoding='utf-8') as f:
    rows = list(csv.DictReader(f))

with open('output.json', 'w', encoding='utf-8') as f:
    json.dump(rows, f, indent=2, ensure_ascii=False)