Flatten JSON

Every nested object is expanded into its own columns, with the path joined by dots, at any depth: {"user": {"address": {"city": "Pune"}}} becomes a user.address.city column. Columns are collected across all records, so a field that only some records have still gets a column, with blank cells for the rest.

This differs from JSON to CSV, which keeps each top-level key as one column and stores nested objects as JSON text. Flatten when you want to filter, sort, or chart nested fields in a spreadsheet. Our guide flattening nested JSON for spreadsheets discusses the trade-offs.

Arrays are not spread into extra rows, which would multiply every other field. An array of simple values is joined with commas into one readable cell, so ["admin", "beta"] becomes admin, beta. An array of objects, such as order lines, is kept as JSON text in a single cell so no structure is lost. An empty array is written as [], and null becomes a blank cell.

If you need one row per array element, for example one row per order line, that is a different operation (an unnest or explode). It is easiest in SQL: load the file into DB in Browser or a tool with JSON functions.

Switch to Nest mode and paste a CSV whose headers use dots. Each dotted header is rebuilt into nested objects, so user.name and user.email end up together under user. Cells that hold JSON arrays or objects, such as the ones Flatten produced, are parsed back into real arrays and objects.

Other values come back as strings, and comma-joined arrays stay as a single string because a comma could just as well be part of the text. Round-tripping is exact for objects and for arrays of objects; arrays of simple values need a split in your own code.