nano-csv2json

Tiny CSV → JSON Lines converter — ~1 KB, zero deps. Single-file, CLI included. Auto-sniffs delimiter; headers optional. Perfect for code-golf, minimal containers, or just for fun.

---

✨ Features

• ✅ CSV → JSONL (one JSON per line)
• ✅ Auto-detect delimiter (,, ;, \t, …)
• ✅ Uses header row if present; --no-header → keys: c1,c2,...
• ✅ Zero dependencies, single tiny file

---

🚀 Usage

# Local (from repo)
python app_min.py data.csv
python app_min.py --no-header data_nohdr.csv

After installing:

# CLI
pip install nano-csv2json
nano-csv2json data.csv > out.jsonl
type data.csv | nano-csv2json         # Windows
cat data.csv | nano-csv2json          # Linux/macOS

Output is JSON Lines (newline-separated JSON objects).

---

🤓 Why so small?

• Uses only the stdlib: csv, json, io — no pandas.
• Auto-detects delimiter with csv.Sniffer() (comma/semicolon/tab…).
• Streams to JSON Lines (one JSON per line) — no in-memory data frames.
• Header row → csv.DictReader; --no-header → c1,c2,... keys via csv.reader.
• Single tiny file + tiny CLI: perfect for scripts, containers, CI.

---

🎉 Fun Ideas

• Pretty-print JSONL

nano-csv2json data.csv | python -m json.tool

• Filter with jq

nano-csv2json data.csv | jq -r 'select(.status=="ok")'

• STDIN piping

curl -s https://example.com/data.csv | nano-csv2json > out.jsonl
# Windows
Invoke-WebRequest https://example.com/data.csv -UseBasicParsing | nano-csv2json > out.jsonl

• Semicolon/TSV auto-sniff

nano-csv2json data_semicolon.csv > out.jsonl
nano-csv2json data.tsv > out.jsonl

• No header → synthetic keys

nano-csv2json --no-header raw.csv | head -1
# {"c1":"val1","c2":"val2",...}

• Compress on the fly

nano-csv2json data.csv | gzip > out.jsonl.gz

• First N records

nano-csv2json data.csv | head -100 > sample.jsonl

---

Tip: Need a single JSON array? Collect lines with jq: nano-csv2json data.csv | jq -s '.' > array.json

---

📜 License

MIT © 2025