JSON to CSV Converter - Online JSON to Excel

1. JSON to CSV in three steps

Aim for a table you can spot-check before dumping the entire dump. Nested arrays explode width if expanded blindly — stringify or normalize upstream when you need multiple columns per array item.

1. Paste JSON and confirm the shape: array of objects is ideal; a single object becomes one row; arrays of primitives become index/value pairs.
2. Enable flatten nested objects when you want address.city-style columns; leave it off to stringify objects/arrays into cells.
3. Pick delimiter and headers, export ten rows for a smoke test, then copy or download the full CSV and open it as UTF-8.

2. Nesting, arrays, and large numbers

Tables are rectangles; JSON is a tree. Flattening joins keys with dots; without flattening, nested values become text cells. Large integers follow a lossless-oriented path so raw literals are preserved when possible, reducing JavaScript Number precision loss. Write the chosen flatten policy into your team notes so exports stay comparable week to week.

• Expand arrays upstream if you truly need multiple columns.
• Export identifiers as text-friendly values when Excel is next.
• Do not flip flatten on and off from muscle memory — decide per schema.

3. Why the formula guard defaults on

Excel and WPS may execute cells that begin with =, +, -, or @. Untrusted CSV then becomes executable content. This tool prefixes a tab by default so those cells display as text. Disable the guard only when the data is trusted and a leading equals sign must survive for a deliberate spreadsheet formula workflow.

4. When not to use JSON→CSV

Irregular deep trees yield sparse wide tables. Relational multi-table data needs splitting or a database load. Tens of megabytes can freeze a tab — browser conversion is for medium debugging loads, not production ETL. Move heavy jobs to scripts or streaming tools.

• Normalize deep trees before tabular export.
• Split detail tables into separate files.
• Use offline ETL for large archives.

5. Import checklist

Export success is not business readiness. Keep the original JSON until the importer accepts the file.

1. Spot-check at least three data rows for nulls and header names.
2. Verify long ID columns did not become scientific notation.
3. Confirm delimiter versus locale (comma vs semicolon) so rows do not collapse into column A.

6. Related tools on this site

Format or sort JSON first when the payload is messy. Prefer JSON→Excel when colleagues need a workbook. Use CSV→JSON for the reverse hop, and CSV→SQL when the destination is a database draft. Keep the chain local and label each artifact with source endpoint and flatten settings.

7. Schema drift, quoting, and column stability

Object arrays whose keys appear only on some rows produce sparse columns when headers are unioned. Downstream SQL loaders may treat missing cells as null — confirm that matches your warehouse expectations. When an API adds a field mid-week, your CSV suddenly gains a column and shifts scripts that assumed positions instead of header names. Prefer header-based imports. If you must feed a system that wants a fixed column list, project the JSON first before export so the CSV shape stays stable. Document flatten depth: address.city is usually welcome; metadata.labels.0.name may be a smell that you should normalize arrays upstream. Fields containing the delimiter, quotes, or newlines must be quoted; this exporter follows that habit. If a downstream loader is quote-blind, pre-clean newlines or choose a rarer delimiter. Always open the CSV in a text editor when an importer fails on a late line number — the real break is often an earlier unclosed quote.

• Import by header name, not by column index.
• Project to a fixed key set when the consumer cannot tolerate drift.
• Treat indexed array columns as a cue to normalize upstream.