SQL Sandbox
The Sandbox is a browser-based SQL editor for querying any dataset on OpenData. Write SQL, see results instantly, and save queries as reusable views.
Getting started
The sandbox needs to know which dataset you want to query. You pick one by searching on the sandbox page itself, or by passing a dataset parameter in the URL:
/sandbox?dataset=provider/dataset
For example, /sandbox?dataset=census/state-populations opens the Census state populations dataset, ready to query.
Finding a dataset slug
Every dataset on OpenData has a two-part identifier: provider/dataset. The provider is the organization that publishes the data (like census, bls, or noaa), and the dataset is the specific table within that provider.
You can find the right slug a few ways:
Search right here
Type a topic below and select a dataset. You'll get the exact slug to use in the sandbox.
Find a dataset slug
Use the search page
The search page lets you browse the full catalog with filters for provider, category, and status. Each dataset card shows the slug in the format provider/dataset.
Use the API
The search endpoint returns dataset metadata including the slug. This is useful if you want to find datasets programmatically.
Search for datasets via API
curl "https://api.tryopendata.ai/v1/search?q=inflation&limit=3" | jqEach result includes provider and slug fields. Combine them as provider/slug to get your dataset identifier.
Writing SQL
Once you've loaded a dataset, the sandbox generates a starter query that selects a few columns. You can edit this freely. Your data is available as a table called data, or you can use the qualified name in double quotes:
-- Short formSELECT * FROM data WHERE year >= 2020-- Qualified formSELECT * FROM "census/state-populations" WHERE year >= 2020
Click any column name in the schema sidebar to insert it at your cursor. The “How do I...” section has templates for common patterns like aggregations, window functions, and date filtering.
For a full reference on supported SQL syntax, functions, and limitations, see the SQL Query documentation.
Builder mode
If you'd rather not write SQL, switch to Builder mode using the toggle in the toolbar. Builder gives you a point-and-click interface for filtering, sorting, selecting columns, and aggregating data. It generates the same query under the hood.
On mobile devices, the sandbox defaults to Builder mode since the SQL editor isn't practical on small screens.
Cross-dataset queries
You can query multiple datasets in a single SQL statement by using their qualified names. Open the sandbox without a dataset parameter and reference tables directly:
SELECTpop.state,pop.population,cpi.value AS price_indexFROM "census/state-populations" popJOIN "bls/cpi-u" cpiON pop.year = cpi.yearWHERE pop.year = 2023LIMIT 20
Saving as a view
Once your query returns results, hit Save (or Cmd+S) to open the save dialog. Give your view a Name (the slug in its URL auto-derives from it), tweak the Slug if you want, and add an optional Description.
The publish toggle controls visibility: left off it saves as a draft only you can see while you iterate, and flipped on it publishes at 1.0.0 and lands on your public profile.
A published view gets its own URL at /@username/views/slug (and a versioned API endpoint), so you can share your curated slice of the data with anyone. The same dialog also handles editing your drafts and forking someone else's view into a copy under your profile.
For the full story on drafts, versioning, forking, and cross-dataset SQL views, see the Views documentation.
URL parameters
The sandbox reads its state from the URL, so you can bookmark or share exact configurations:
| Parameter | Description | Example |
|---|---|---|
dataset | Dataset to load | census/state-populations |
mode | sql or builder | sql |
view | Load a saved view by name | top-states-2023 |
autorun | Run the query immediately on load | 1 |
Next steps
- SQL Query reference for the full syntax, allowed functions, and limits
- Views to learn about saving and sharing query results
- Querying guide for filtering, sorting, and aggregation via the REST API