CSV to JSON converter

CSV has no types: every field is text. With type detection on, the converter turns a value into a JSON number, boolean or null only when that loses nothing. Everything else stays a string.

Types are also checked per column. If one value in a column has to stay text, such as the zip code 04050, every number in that column stays text too. So a zip column gives "04050" and "78701", never a mix of strings and numbers that breaks code expecting one type.

Many converters, including Papa Parse with dynamicTyping, turn 007 into 7. That breaks zip codes, SKUs and account numbers, which is why this tool keeps them as text.

With "Keyed by first column", rows that share a key overwrite each other; the tool tells you how many did.

JavaScript with Papa Parse, the parser this page uses. Splitting on commas by hand breaks on quoted fields like "Desk lamp, brass", so use a real parser:

import Papa from 'papaparse';

const { data, errors } = Papa.parse(csvText, {
  header: true,          // first row = keys
  skipEmptyLines: true,
  dynamicTyping: true,   // "42" -> 42, "true" -> true (also "007" -> 7)
});
const json = JSON.stringify(data, null, 2);

Python with the standard library. utf-8-sig drops the byte order mark Excel adds to UTF-8 CSV files, so your first key is not \ufeffsku:

import csv, json

with open("products.csv", newline="", encoding="utf-8-sig") as f:
    rows = list(csv.DictReader(f))      # values stay strings

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

pandas infers types per column, so pass dtype=str for columns with leading zeros:

import pandas as pd

df = pd.read_csv("products.csv", dtype={"zip": str})   # keep leading zeros
df.to_json("products.json", orient="records", indent=2, force_ascii=False)

Miller on the command line converts between CSV, TSV, JSON and JSON Lines:

# Miller: CSV in, JSON out
mlr --icsv --ojson cat products.csv > products.json

# One JSON object per line instead
mlr --icsv --ojsonl cat products.csv > products.jsonl

• A field with the delimiter, a double quote or a line break must be wrapped in double quotes (RFC 4180).
• A double quote inside a quoted field is written twice: "Cable tray ""XL""" reads as Cable tray "XL".
• Line breaks may be CRLF (Windows, the RFC) or LF (Unix); both work, and blank lines are skipped.
• The header row is optional. Without it, keys are column_1, column_2 and so on; blank headers get the same names and repeated ones get _2, _3.
• Rows with more or fewer fields than the header are listed with their row number, so you can find the stray comma or quote.

Need the reverse? JSON to CSV flattens nested JSON back into columns. To check the JSON you get, use the JSON formatter, and for config files try JSON to YAML. Building on top of data? See database tools on DevHunt.