Data science: Convert CSV to JSON lines

Many tools want JSON rather than CSV. Convert a CSV table into one JSON object per line, with every value given its proper type.

What to build

A program that converts CSV rows into typed JSON objects.

Requirements

  • Read the lines with the csv module. Convert each value: an empty field becomes null; a whole number (optional minus sign, then digits) becomes an integer; anything else that float() accepts becomes a decimal; true and false (lower case) become booleans; anything else stays a string.
  • Print each data row as json.dumps of a dictionary with the header's keys in order, then rows converted: N.

Input

Line 1: n, the number of CSV lines including the header. Then the n lines.

Output

One JSON object per data row, then the count.

Constraints

1 ≤ n ≤ 1000; values never contain line breaks.

Example

Input:

3
id,name,score,passed
1,Asha,88.5,true
2,"Lee, Jr.",,false

Output:

{"id": 1, "name": "Asha", "score": 88.5, "passed": true}
{"id": 2, "name": "Lee, Jr.", "score": null, "passed": false}
rows converted: 2

Why it works

Converting each field in a fixed order of checks gives every value its natural type once, at the boundary, so the JSON consumers get numbers, booleans and nulls instead of strings they would have to guess about.

Examples

Example 1 · Input
3
id,name,score,passed
1,Asha,88.5,true
2,"Lee, Jr.",,false
Expected output
{"id": 1, "name": "Asha", "score": 88.5, "passed": true}
{"id": 2, "name": "Lee, Jr.", "score": null, "passed": false}
rows converted: 2
Example 2 · Input
2
x
007
Expected output
{"x": 7}
rows converted: 1

Prompts inside input() are ignored when checking, so input("n: ") is fine.

Hints

Hint 1

Write convert(text) that tries, in order: empty, whole number, decimal, true/false, and otherwise keeps the text.

Hint 2

Build each row as a dictionary in header order; json.dumps keeps that order.

Solution

Show one correct solution (try it yourself first)
import csv
import json
import re

n = int(input())
rows = list(csv.reader([input() for _ in range(n)]))
header, data = rows[0], rows[1:]


def convert(text):
    if text == "":
        return None
    if re.fullmatch(r"-?\d+", text):
        return int(text)
    try:
        return float(text)
    except ValueError:
        pass
    if text in ("true", "false"):
        return text == "true"
    return text


for row in data:
    print(json.dumps({key: convert(value) for key, value in zip(header, row)}))
print("rows converted:", len(data))

Lesson: Reading CSV and JSON data →