JSON arrays are frequently ragged

An API response omits fields that are null, a database export adds a new column halfway through the records, and a form collection grows optional fields over time. A table needs fixed columns, but your JSON often doesn't — different objects in the same array can have completely different keys. This converter reconciles that mismatch by building a table where the column set is the union of every key across every object, not just the keys of the first one. One record that introduces a new key adds a column for the entire table.

Column order follows first appearance

When objects introduce new keys in different positions, how should the table render them? This converter walks the array in sequence and appends each new key to the column list when it is first encountered. So if the first object has name and age, the second adds city, and the third repeats keys already seen, the final column order is name, age, city — not alphabetical, and not the order of whichever object happens to be widest.

The practical consequence: the shape of your table depends on the order of your data. Sort the same array differently and the columns can come out in a different order, even though no value has changed. If you need a stable column order across runs, sort or normalise the records before converting.

Missing keys and nested values need attention

When an object lacks a key that appears elsewhere, its cell is empty — not the text undefined, not null, just blank. If a key is explicitly null in the JSON, that also becomes an empty cell, so a missing key and an explicit null are indistinguishable in the output. That distinction often matters — "we never collected this" is not the same fact as "we collected it and it was nothing" — so if it matters to you, substitute a placeholder before converting.

Objects and arrays inside cells are serialised inline with JSON.stringify, so a nested address becomes {"city":"London","country":"UK"} right in the cell rather than being spread across extra columns. This keeps the data visible and lossless, but can make one column far wider than the rest. If that hurts readability, flatten the structure before converting.

Input must be a non-empty JSON array of plain objects

Three error messages tell you exactly what went wrong. Malformed JSON gives Invalid JSON: followed by the parser's own message, which usually names the position of the problem. A single top-level object, or an empty array, gives Expected a non-empty JSON array of objects — this is by far the most common mistake, and the fix is simply to wrap your object in square brackets. An array holding a number, string, null or nested array rather than an object gives Every array item must be a plain object.

For the mechanics of the Markdown table format itself — pipe escaping, padding and the separator row — see the CSV to Markdown Table Converter.