An order ID such as 0012 is a label. Converting it to the number 12 changes the identifier. CSV/JSON Converter protects leading-zero strings, but you should still choose the conversion settings deliberately and inspect the result.
Choose the source and turn inference off when needed
Choose Delimited text to JSON and paste the table, or select one file. A chosen file takes precedence over the text box. Set the delimiter explicitly when you know it, and confirm whether the first row contains column names.
The current Infer numbers, booleans and literal null option starts enabled. It keeps leading-zero values and long numeric strings as text, but converts eligible numeric values and exact lowercase true, false and null. Turn it off when every cell must remain a string, including quantities and words that happen to match those literals.
Quoting a value in CSV is a way to represent its contents; it is not a portable declaration of that column's data type. Use the converter's inference setting and inspect the JSON types.
Try an inventory example
sku,quantity,enabled
0012,3,true
0107,12,false
With inference off and Array of objects, the first row becomes:
{"sku":"0012","quantity":"3","enabled":"true"}
With inference on, its SKU remains "0012", while quantity becomes 3 and enabled becomes true. Choose the representation the receiving system expects. If it needs only selected columns converted, preserve strings here and apply field-specific typing in that system.
Inspect columns and output shape
Unique, meaningful header names make object properties easier to map. Repeated headers can be renamed by the reader; inspect the actual keys. CSV Viewer helps inspect a source, and CSV Cleaner can prepare headers before conversion.
Array of arrays with header row produces a different structure from an array of objects. Object keyed by one unique column requires nonblank, unique keys and refuses duplicates instead of overwriting a row. Match the shape to the importer's specification rather than choosing the shortest output.
Quoted commas and line breaks belong inside a cell when the file quotes them correctly. Review delimiter and extra-field warnings. RFC 4180 documents common CSV quoting conventions; files using other delimiters need the corresponding settings.
Check spreadsheet sources before converting
For .xlsx, choose Excel workbook to JSON and select the worksheet by name or position. An empty selection uses the first visible sheet. Recognized date and time formats become ISO values using the workbook's date system; unsupported numeric formats can remain numeric values.
Saved formula results are not recalculated. The advanced formula option can use cached results or return formula text. A missing saved result can prevent conversion. Also, a numeric cell displayed as 0012 through a number format has not necessarily stored those leading zeros as text. Prepare a real text identifier column in the workbook before conversion when the spelling must survive.
Verify the download
Compare data-row and column counts, then spot-check identifiers, blank values and types in the JSON. A formatter's total object count is not a reliable substitute for the top-level row count when structures differ.
One input file can be up to 20 MiB; text input allows 10,000,000 characters and conversions allow 200,000 rows. Workbook expansion has additional safety limits. Keep the source and converted file separately. Once an earlier spreadsheet or export has removed a leading zero, this converter cannot infer the missing digit.