CSV and JSON represent different things. CSV stores rows of text cells. JSON can express nested objects, arrays, numbers, booleans and null. Conversion needs decisions about those differences; a successful download alone does not prove the values still mean the same thing.
Start with a small representative sample
Before converting an export, make a copy containing a normal row, a missing value, a comma inside a value, and a code with a leading zero. Avoid using production secrets as test data. This sample is enough to check the browser tool's main parsing behavior:
code,name,city
0012,"Ada, A.",Delhi
0013,Lin,
Paste it into CSV to JSON. The first row becomes the property names. The code stays the string "0012", the quoted comma stays inside the name, and the missing city becomes an empty string. The current converter deliberately keeps cell values as strings; it does not infer dates, numbers or booleans.
Protect identifiers from automatic type conversion
A postal code, telephone number or account identifier is text even when it contains only digits. Converting 0012 to the number 12 removes information. Large identifiers can also exceed JavaScript's exact integer range if treated as numbers.
Keep codes as strings in the source JSON and check the importing application too. A spreadsheet may interpret a text-looking CSV cell as a number unless you choose the column type during import. Quotation in CSV controls cell boundaries; it is not a universal instruction to every spreadsheet to preserve a text type.
If a downstream API needs a numeric quantity, convert that specific field after validation. A blanket rule that turns every digit-only cell into a number can damage the identifiers in the same record.
Confirm delimiters and headers
This site's CSV parser expects commas. It does not automatically select semicolons or tabs. An export that appears in a single JSON property may be using the wrong delimiter. Change the export settings or use a parser with an explicit delimiter setting.
Use a nonempty, unique header for each column. This converter rejects blank or duplicate headers instead of inventing column names or overwriting values. For example, code,code is rejected even if both cells contain the same value. Rename the headers in the source before converting.
Every data record must have exactly as many fields as the header. Missing or extra fields produce an error; the tool does not silently drop extra cells or pad a short record. Under a three-column header, 0013,Lin, supplies an intentionally empty third field, while 0013,Lin is missing that field. Correct the source or its export settings before retrying.
Distinguish an empty record from a blank line
Empty strings and whitespace are data. This two-column example retains both records and preserves the two spaces in the second name:
name,note
"",""
" ",""
The result is [{"name":"","note":""},{"name":" ","note":""}], displayed with formatting. A completely empty physical line outside a quoted field is skipped. To keep an intentional empty record, supply all its fields with delimiters or quoted empty strings. Do not use a blank line as a placeholder for a row you need to preserve.
A quoted line break inside a cell belongs to that record and is preserved. A missing closing quote, an unescaped quote inside an unquoted field, or text after a closing quote is rejected. These are reasons to use a CSV parser instead of splitting the file with a plain text-line tool.
Understand JSON to CSV flattening
JSON to CSV accepts an object or an array of objects. It builds a header list from keys across the objects in their first-seen order. Missing properties and null values become empty cells. That means a round trip cannot distinguish an absent property, null and an empty string just from the resulting CSV.
Nested arrays and objects become a JSON string inside one cell. They are not expanded into multiple columns. For example, {"address":{"city":"Delhi"}} creates an address cell containing serialized JSON, rather than an address.city column. If your destination requires flat columns, flatten the object deliberately before conversion.
Commas, double quotes and line breaks in text require CSV quoting. The converter wraps such cells in double quotes and doubles embedded quotation marks. Open the result through a CSV import flow that understands those rules, rather than judging it by how many commas appear in a text editor.
Treat spreadsheet formulas separately
A value beginning with = can be interpreted as a formula by some spreadsheet import workflows. This converter serializes values; it does not provide a spreadsheet formula-sanitization mode. If data comes from other people, import suspect columns explicitly as text and follow the destination application's controls before opening the file.
Do not remove the leading character automatically if you need an exact archival copy. Keep a raw export and a separately prepared spreadsheet import when those purposes differ.
Verify before replacing the original
Compare row counts, headers and the sample values after import. Check a code with leading zeros, a quoted comma, an empty field and any nested value. Use JSON Formatter to inspect structure, while remembering that syntactically valid JSON can still contain the wrong business data.
The examples above were checked against this site's converter source on October 16, 2026. For the common quoting conventions and their interoperability limits, see RFC 4180. For numeric precision, see MDN's Number.MAX_SAFE_INTEGER reference.