Show HN: Transform a CSV into a JSON and vice versa

Looks a bit like adware IMO. The library appears to be drenched in analytics. Dependencies include:

https://www.npmjs.com/package/web-vitals/v/0.1.0 https://www.npmjs.com/package/@fingerprintjs/fingerprintjs

Harvesting user's data, most likely...

At work we work an a Python Library to do this, and much more:

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

Source: https://github.com/OpenDataServices/flatten-tool

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

It converts JSON to CSV and vice versa but also Spreadsheet files, XML ...

It has recently had some work to make it memory efficient for large files.

Work, BTW, is an Open Data Workers Co-op working on data and standards. We use this tool a lot directly, but also as a library in other tools. https://dataquality.threesixtygiving.org/ for instance - this is a website that checks data against the 360 Giving Data Standard [ https://www.threesixtygiving.org/ ].

Slightly OT: I've realized that CSVs are dramatically more information-dense than the equivalent JSON, and actually make a pretty reasonable API response format if your dataset is large and fits into the tabular shape. They can be a fraction of the size, mainly because keys aren't duplicated for every item.

I wrote my own converter a few years ago, then ended up needing it again last week. It’s one of those things you don’t always need but it’s handy to have when you do. https://github.com/baltimore-sun-data/csv2json

I recently heard about a tool called Miller that helps convert between JSON and CSV among other formats: https://github.com/johnkerl/miller

  mlr --c2j cat documents.csv > documents.jsonl

CSV is more of a rumor than a standard, plus JSON can have a tree structure. It is a fun idea to think about and may be useful in some narrow cases, but will fail in almost all but those most trivial of structures.

jq's stream and fromstream functions can be used to flatten and unflatten JSON. I use it all the time at work for POs who want to see data in Excel

https://jqplay.org/s/ub-WvXCcPn

... from there it's just a row->column rotation to CSV.

CSV<->JSON is fundamentally an unsolvable problem because of mismatch in data hierarchies among them.

Plus you have the type looseness for both and lack of standards for CSV.

A trivial 2-D case is handled well by Python library such as Pandas. Here OP could be an alternative.

When I say trivial I mean flat 2 dimensional data, such as you would get from Mockaroo or similar source.

However in real life - data is messy.

As you get into 3,4 and deeper hierarchies on JSON you can't really translate that into nice flat 2d CSV.

Then you have missing keys, mixed up types and you end up rolling you own hand written converters.

I have a couple of open source CLI tools to do this: - https://github.com/tyleradams/json-toolkit/blob/master/csv-t... - https://github.com/tyleradams/json-toolkit/blob/master/json-...

Wouldn't this need to allow upload/download of CSV to really meet the spirit of the title? Or maybe replace the references of CSV with "HTML Table"?

If anyone is looking at a free tool to quickly edit CSV data in an Excel-like editor, see https://editcsvonline.com

Built something similar on codepen quite a few years ago. Not sure where I came up with the format, seems a bit wild looking at the v. nice dot notation used here, but possibly more useful/efficient for variable data models, also takes into account data types:

https://codepen.io/theprojectsomething/pen/OwppWW

Note: click the Toggle Info to read the "spec" (groan) :)

I run a software company and we have a challenge when it comes to these types of conversion tools.

If there is any data that is a) not publicly accessible or b) contains personal information, I cannot authorise the use of a web based third party tool. There is just too much risk that some bad actor uses this as a method to soak up data.

I would love to verify /validate that all of the processing is local and have some way to certify if this hasn't changed.

It would make it much easier for users to visually parse the JSON section if you added `font-family: monospace` to the textarea element

Nice, I like the look of the editable table.

Shameless plug: a similar solution, working all client side, not imposing to use a key as first column, and with options regarding CSV format.

https://mango-is.com/tools/csv-to-json/

this can turn an HTML table into JSON with the option to download the JSON, or it can turn JSON into an HTML table with no option to download a CSV -- where does CSV come into this?

also, clearly javascript is a bit too ambitious for the job when e.g. PHP could provide the intended functionality with two lines of code: foreach($arrays as $values){echo implode(',', $values) . "\n";} echo json_encode($arrays);

also, CSV is more for storing rigidly-structured uniform columns and rows, whereas JSON is more for storing loosely-structured varying objects, otherwise you're redeclaring column headings in every array, which wouldn't make much difference for gzipped transport but still wasteful and verbose nonetheless. column headings are usually the first line of a CSV

Why doesn't the library transform the csv into a an array of json objects?

FYI, the view on small devices is pretty bad - the demo json output is almost unreadable without an awkward pinch + scroll. Compare this view to the same content on the Readme via GitHub.

Shameless plug - I wrote convertcsv.com and it supports about everything you can think of as far as format conversions. JSON, XML, YAML, JSON Lines, Fixed Width, ...

You can do this in Powershell

    cat file.csv | ConvertFrom-Csv | ConvertTo-Json

`jq` can transform CSV to JSON and vice-versa, especially for simple/naive data where simply splitting on `,` is good enough - and where you aren't too bothered by types (e.g. if you don't mind numbers ending up as strings).

First attempt is to simply read each line in as raw and split on `,` - sort of does the job of, but it isn't the array of arrays that you might expect:

    $ echo -e "foo,bar,quux\n1,2,3\n4,5,6\n7,8,9" > foo.csv

    $ jq -cR 'split(",")' foo.csv
    ["foo","bar","quux"]
    ["1","2","3"]
    ["4","5","6"]
    ["7","8","9"]

   $ jq -R 'split(",")' foo.csv | jq -cs
   [["foo","bar","quux"],["1","2","3"],["4","5","6"],["7","8","9"]]

   $ jq -cR 'split(",")' foo.csv | jq -csf csv2json.jq
   [{"foo":1,"bar":2,"quux":3},
    {"foo":4,"bar":5,"quux":6},
    {"foo":7,"bar":8,"quux":9}]

In the reverse direction there's a builtin `@csv` format string. This can be use with the second example above to say "turn each array into a CSV row" like so:

   $ jq -R  'split(",")' foo.csv | jq -sr '.[]|@csv'
   "foo","bar","quux"
   "1","2","3"
   "4","5","6"
   "7","8","9"

    $ jq -cR 'split(",")' foo.csv | jq -csf csv2json.jq | \
    > jq -r '.[]|[.foo,.bar,.quux]|@csv'
    1,2,3
    4,5,6
    7,8,9

[0] https://github.com/stedolan/jq/wiki/Cookbook#convert-a-csv-f...

I'm sorry, but why is this a website?

How do you actually load CSVs into it?

Can this handle uploading csvs?

You should use outline instead of border in your cell css.