How to Convert CSV to JSON Step by Step

CSV stores data as rows of comma-separated values, while JSON stores the same data as an array of objects. Converting one to the other is mostly mechanical, but the details — headers, quoting and types — decide whether the result is clean or broken.

The mapping: headers become keys

Each CSV column header becomes a JSON property name, and each data row becomes one object in an array. Take this input:

name,age,city
Ada,36,London
Linus,54,Helsinki

The converter produces:

[
  { "name": "Ada", "age": "36", "city": "London" },
  { "name": "Linus", "age": "54", "city": "Helsinki" }
]

Every value is a string by default, because CSV has no concept of types. Turning "age": "36" into the number 36 is a deliberate extra step.

Step by step

1. Check the delimiter. Most files use commas, but some use semicolons, tabs or pipes. Detect it before splitting.
2. Read the header row. Trim extra spaces so first name becomes the key first name, not first name.
3. Split each row into fields. Respect quotes: a field wrapped in double quotes may itself contain commas.
4. Build an object per row. Zip the header names with the row values.
5. Collect the objects into an array and serialise to JSON.

Quoted fields and commas

The trickiest part of CSV is that a value can contain the delimiter. This row:

name,address
Smith,"12 Oak Street, Springfield"

has two fields, not three, because the comma sits inside quotes. A naive split on commas would create a phantom column. The rule inside quotes: commas are literal, and a doubled quote ("") represents one real quote character.

Missing values and ragged rows

Real files are rarely tidy. A row may have fewer fields than the header (a missing value) or more (an unescaped delimiter). Decide up front how to handle a short row — usually by filling the gap with null or an empty string — so your JSON objects keep a consistent shape. Consistency matters because most consumers expect every object to have the same keys.

Common pitfalls

Choosing the right tool

For a one-off job, a browser converter is the fastest route. For a repeatable pipeline, use a library so the parsing rules stay consistent. Either way, always check a few rows by eye after converting; CSV's loose rules mean bad data rarely announces itself.

Paste your CSV into our CSV to JSON converter to get a structured array with quoted fields handled automatically, and read JSON vs CSV to decide which format to keep as your source of truth. If your data contains URL-encoded values, the URL encoding guide explains how to read them.