I tend to use python (usually in a jupyter notebook) and pandas. Lots of experimentation with pandas' json_normalize() function. https://towardsdatascience.com/all-pandas-json-normalize-you...

then just call to_csv() on the dataframe. (edited to add to comment on exporting to CSV as per the original question).

I also use jupyter for this, but my goto is tablib. for anyone who hasn't used it, it's super easy to switch between tabular data formats. you create an instance of their Dataset class, then assign your data to the appropriate property, and all of the other properties are your data in the respective format

for instance:

  from tablib import Dataset

  json_array_of_objects = '[{"header": "data1"}, {"header": "data2"}]'
  ds = Dataset()
  ds.json = json_array_of_objects
  
  ds.csv # data formatted as a csv
  ds.xlsx # excel, only useful on a binary read or write
  ds.dict # list of dictionaries
  ds.json # list of dictionaries converted to json
  ds.jira # table formatted for jiras markup
  ds.html # html table
  # and more

they used to vendorize dependencies, so everything worked out of the box, but now some features need to be installed specifically, or do pip install tablib[all], which is kind of annoying. I suspect they started doing it when they included support for pandas dataframes, because they didn't want to vendorize all of pandas. or force it to install as a requirement.

In addition to the already mentioned jq, there's https://github.com/jehiah/json2csv

Try this: https://flatterer.opendata.coop/

There is no binary yet but there is a python CLI and library, even though it is written in rust.

It is the only tool that I know that deals with nested JSON and converts it into relational tables.

Here is a notebook of the python library usage.

https://deepnote.com/@david-raznick/Flatterer-Demo-FWeGccp_Q...

Try https://github.com/tomnomnom/gron

I mostly go from csv to csv but nushell handles table data in files really well [1]. I've done a little but of json manipulation with it but not a ton, so I can't say it's a silver bullet.

1. https://www.nushell.sh/book/

Here's a Python library with a CLI:

Pypi: https://pypi.org/project/flattentool/

Docs: https://flatten-tool.readthedocs.io/en/latest/

It's maintained by Open Data Services Coop, where we use it as a component in several of our web & data pipeline tools for working with data that is published in a Data Standard.

https://github.com/stedolan/jq can manipulate arrays of objects and format as CSV

Of course you can do it with http://brackit.io.

What exactly is the "transformation" you envision?

Something like this (very simplified):

  let $array := [{"foo":0,"bar":"tztz"},{"foo":"hello","bar":null},{"foo":true,"bar":"yes"}]
  let $value := for $object in $array
                return
                  let $fields := bit:fields($object)
                  let $len := bit:len($fields)
  
                  for $field at $pos in $fields
                  return if ($pos < $len) then (
                    $object=>$field || ","
                  ) else (
                    $object=>$field || "\n"
                  )
  return string-join($value,"")

will output:

  0,tztz
  hello,null
  true,yes

You could do something like this in pure python without the json loading boilerplate with jello[0]. An interactive TUI for jello called jellex[1} is also available. (I am the author)

[0] https://github.com/kellyjonbrazil/jello

[1] https://github.com/kellyjonbrazil/jellex

JSON loading boilerplate? Is it also a compiler with sophisticated optimizations during compile time? :)

Of course you can simply create a module with something like the above snippet as a function for reuse. Or I can add a built-in function for flattening like this or as in pandas.

If you use JavaScript-> papaparse or like other have pointed out straight forward to do in python, bash, Perl, etc…

Excel can do this.

jq