Data Studio

One tree, six formats

Every conversion parses the source into a single typed tree and then writes that tree back out in the target format. That is what makes the round trips honest: a number stays a number, nesting survives, and object keys keep the order they were written in rather than being quietly alphabetised. It is also why adding a format is one parser and one serialiser, not a new special case for every pair.

What each format is good at

JSON

The default interchange format. Keeps nesting and types intact.

YAML

Config-friendly and comment-friendly, though comments are not carried over.

TOML

Needs a table at the root. Nulls are dropped — TOML has no such value.

CSV

Spreadsheet-ready. Needs an array of objects; nested values become JSON cells.

TSV

CSV with tabs. Safer when your data is full of commas.

JSONL

One document per line. The format log pipelines and LLM datasets use.

The awkward edges, handled on purpose

• CSV has no types. A cell only becomes a number when printing that number reproduces the original text exactly, so 007 stays a string while 12 does not.
• CSV has no nesting. Writing a table from nested data keeps the nested values as JSON inside the cell rather than inventing a flattening scheme you would then have to reverse.
• TOML has no null. Keys whose value is null are dropped instead of failing the whole document, and a source that is not an object at the root is rejected outright — TOML needs a table there.
• Comments are not data. YAML and TOML comments live outside the tree, so they do not survive a conversion. Nothing can carry them across.

Frequently asked questions

How do I convert JSON to YAML, or YAML to JSON?

Paste or drop the document, confirm the detected input format, and pick the output format. Both directions go through the same typed tree, so nesting, numbers and booleans survive the trip.

Does converting CSV to JSON keep my leading zeros?

Yes. A cell is only turned into a number when printing that number back reproduces the original text exactly, so an ID like 007 stays the string "007" while 12 becomes the number 12.

Is the key order of my JSON preserved?

Yes. Object keys keep document order rather than being alphabetised, so the output stays diffable against your source file.

Why did my conversion to TOML fail?

TOML needs a table at the top level, so an array or a bare value cannot be represented. TOML also has no null: keys whose value is null are dropped rather than failing the whole document.

What is JSONL, and when should I use it?

JSONL — also called NDJSON — puts one compact JSON document on each line. It is what log pipelines and machine-learning datasets use, because a reader can stream it a record at a time instead of parsing the whole file first.