- Home
- Tools
- Developer Tools
- JSON to CSV Converter
- Convert CSV to JSON
Convert CSV to JSON
Turn CSV into a JSON array of objects. The delimiter is detected, quoted fields are read properly and numbers become numbers.
JSON
How a field gets its type
A CSV file holds text and nothing else. Whether 42 is a number or a house number is not in the file, so the converter has to decide, and it is careful about it. One file with a few awkward fields:
name,joined,member,zip,note "Jansen, A.",2024-03-01,true,0123,"said ""hi""" Bob,2025-11-22,FALSE,90210,
[
{
"name": "Jansen, A.",
"joined": "2024-03-01",
"member": true,
"zip": "0123",
"note": "said \"hi\""
},
{
"name": "Bob",
"joined": "2025-11-22",
"member": false,
"zip": 90210,
"note": ""
}
]What a field turns into
| In the CSV | In the JSON | Why |
|---|---|---|
In the CSV42 | In the JSON42 | Written exactly like a JSON number. |
In the CSV-1.50 | In the JSON-1.50 | A number too. Its digits are kept as they are. |
In the CSV007 | In the JSON"007" | A leading zero means a code, not a quantity. |
In the CSV+31612345678 | In the JSON"+31612345678" | A plus sign is not part of a JSON number, and a phone number is not one either. |
In the CSV1,5 | In the JSON"1,5" | A decimal comma is not how JSON writes numbers. |
In the CSV12345678901234567890 | In the JSON"12345678901234567890" | More digits than JavaScript holds exactly, so it stays text and keeps them all. |
In the CSVTRUE | In the JSONtrue | true and false are recognised in any capitalisation. |
In the CSVnull | In the JSONnull | Only in lowercase. |
In the CSVNull | In the JSON"Null" | Not lowercase, so it stays text. It is also a surname. |
| In the CSV(empty) | In the JSON"" | An empty field is an empty text, not null. |
Fields in quotes
Quotes are how a field holds the delimiter, a line break or a quote of its own. They are taken off when the file is read.
| Written in the file | The value | Why |
|---|---|---|
Written in the file"Jansen, A." | The valueJansen, A. | Quotes keep the comma inside one field. |
Written in the file"said ""hi""" | The valuesaid "hi" | Two quotes in a row stand for one. |
Written in the file" padded " | The value padded | Spaces inside the quotes belong to the value. |
Written in the file6" nails | The value6" nails | A quote in the middle of a field without quotes is just a quote. |
Good to know
The decision is made per field, not per column. A column of postcodes can come out as text in one row and as a number in the next, as 0123 and 90210 do above. If that matters, switch the conversion off and every value stays text.
Dates stay text, because JSON has no type for them.
The delimiter is taken from the first twenty rows: comma, semicolon, tab or pipe, whichever gives every row the same number of fields. A first line such as sep=; is obeyed and left out of the result.
Column names come from the first row. An empty name becomes column_3, after its position, and a name that occurs twice gets _2 behind it. A row with too few fields is filled up with empty text.
With Nest names with dots switched on, a column called address.city becomes an object and lines.0.sku a list.
The JSON to CSV converter page has more on delimiters, Excel and cells that start with an equals sign.