# Data Explorer

> Source: https://docs.clonepartner.com/ui/data-explorer/

Data Explorer is the interactive query surface for stored connectors. It discovers resources through the connector, sends queries to the server, and keeps query state in the URL so a view can be revisited.

## Start a query

1. Open **Data Explorer**.
2. Select a connector.
3. Select a resource from the sidebar.
4. Choose **Visual** or **Raw** mode.
5. Set filters, sort, columns, and a row limit.
6. Select **Run** or **Send**.

Selecting a resource does not imply a write. Connectors with a request builder wait for an explicit send.

## Visual mode

Visual mode builds connector-native arguments from structured controls. Depending on the connector, you can:

- find or count rows;
- add typed filters;
- choose server-side sorting;
- select fields;
- limit and page results.

MongoDB filters preserve ObjectId, date, number, boolean, null, and array types. SQL connector filters use the runtime's JSON filter dialect.

## Raw mode

Raw mode exposes connector-specific JSON, SQL, OData, or method arguments. Use it when a query cannot be expressed by the visual builder.

Raw mode is not a bypass around authorization. Mutating methods remain limited to admins and superadmins, and the server validates allowed explorer methods.

:::callout{type="warning"}
Treat raw write queries as production changes. Preview a read query first, narrow the filter, and confirm the selected connector and environment.
:::

## Results

Tabular results can be viewed as a table or JSON and transformed for display with JSONata. You can copy rows and export the current page as CSV or NDJSON.

Where supported, **Export all** streams matching rows from the server instead of loading the full dataset in the browser. Cancel a long export from the same menu.

## Pagination

Explorer list operations request one page at a time. Use next and previous cursor controls where a connector supplies them. Row limit is not a substitute for a complete export.

S3 browsing is hierarchical and paged. Small supported objects can preview; larger or binary objects remain download-only.

## Truto and catalog-backed connectors

For connected-account connectors:

1. choose the connected account;
2. select a unified or proxy resource;
3. select a supported method;
4. add documented query fields;
5. send the request.

The builder exposes resource ID and body fields only when required by the method. Custom API mode allows an explicit HTTP verb and path for provider operations not represented by a named resource.

Write methods and row edit/delete actions are shown only to roles allowed to mutate.

## Team presets

Presets save connector, resource, mode, filters, columns, limit, result view, and optional JSONata state. They are visible to the team.

Use parameters for values that change each time:

- a free-text string;
- a Mongo ObjectId;
- a searchable record selected from a Mongo collection.

Parameterized raw presets require valid JSON and are intended for JSON-based connector queries. SQL and OData raw text cannot use preset parameters.

Open **Presets → Save team preset** only when the query is useful to others. Give it a purpose-based name and avoid embedding personal or secret values.

See [Use Data Explorer presets in practice](/guides/data-explorer-presets-in-practice) for validation rules and an end-to-end example.

## Safe exploration

- Begin with the smallest useful limit.
- Prefer read-only connector accounts for investigation.
- Verify the connector name and environment before writes.
- Keep exports in approved storage.
- Do not use browser-side filtering as a substitute for a server query.
- Add indexes before enabling frequent sort or filter paths over large stores.