> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/metabase/metabase/llms.txt
> Use this file to discover all available pages before exploring further.

# Asking questions

> Learn how to create and save questions in Metabase to analyze your data using the query builder or SQL editor.

Questions in Metabase are queries, their results, and their visualization. Questions are the basic analytical unit in Metabase - you can think of them as saved queries that you can display as tables or charts.

## What is a question?

A question is composed of three parts:

* **The query**: The data you want to retrieve from your database
* **The results**: The data returned by your query
* **The visualization**: How you choose to display those results (table, chart, etc.)

You can organize questions into collections and dashboards, embed them, share links to them, export their results, and set up alerts when results are interesting.

## Creating a new question

There are several ways to create a question in Metabase:

<Tabs>
  <Tab title="From scratch">
    Click **+ New** in the main navigation and choose how you want to query your data:

    * **Question**: Opens the graphical query builder
    * **SQL query**: Opens the native SQL editor

    The query builder is great for most questions and doesn't require SQL knowledge. The SQL editor gives you full control for complex queries.
  </Tab>

  <Tab title="From existing question">
    You can build on top of any existing question without modifying the original:

    1. Open the question you want to build from
    2. Click the three-dot menu
    3. Select **Duplicate** or use it as a data source for a new question

    <Note>Some question types can't be used as source data, including Druid questions, MongoDB questions, and questions using cumulative aggregations.</Note>
  </Tab>

  <Tab title="With Metabot">
    Use natural language to create questions with AI assistance:

    1. Click **+ New** > **Question**
    2. Select **Ask Metabot**
    3. Describe what you want to know about your data

    Metabot can generate both query builder questions and SQL queries based on your request.
  </Tab>
</Tabs>

## Choosing your data source

When creating a question, you'll first select where your data comes from:

<CardGroup cols={2}>
  <Card title="Tables" icon="table">
    Query data directly from database tables. Select your database, then choose the specific table you want to analyze.
  </Card>

  <Card title="Models" icon="diagram-project">
    Start from a curated dataset. Models are specially prepared questions that serve as good starting points, similar to views in SQL.
  </Card>

  <Card title="Saved questions" icon="bookmark">
    Build on existing questions. Use any saved question's results as the foundation for a new analysis.
  </Card>

  <Card title="Metrics" icon="chart-line">
    Begin with pre-defined calculations. Metrics are standardized measures like revenue or active users defined by your team.
  </Card>
</CardGroup>

## Building your question

<Steps>
  <Step title="Select your data source">
    Choose a table, model, saved question, or metric as your starting point.
  </Step>

  <Step title="Filter your data">
    Narrow down the data by adding filters. You can filter by date ranges, categories, numeric values, and more.
  </Step>

  <Step title="Summarize and group">
    Calculate metrics like counts, sums, and averages. Group by dimensions like time, location, or category to break down your metrics.
  </Step>

  <Step title="Visualize your results">
    Choose the best chart type for your data. Metabase will suggest appropriate visualizations, but you can customize them.
  </Step>
</Steps>

## Saving your question

Once you've built your query and visualized the results, you can save your question in two ways:

<Tabs>
  <Tab title="Save to collection">
    **Best for reusable questions**

    Questions saved to collections can be:

    * Added to multiple dashboards
    * Found through search
    * Shared with team members
    * Used as data sources for other questions

    To save to a collection, you'll need curate access to that collection.
  </Tab>

  <Tab title="Save to dashboard">
    **Best for dashboard-specific questions**

    Questions saved directly to dashboards:

    * Only appear in that specific dashboard
    * Can't be reused in other dashboards
    * Are arranged and saved when you save the dashboard

    This is useful for questions that are highly specific to one dashboard's context.
  </Tab>
</Tabs>

<Info>
  When saving a question, you can add a name and description. The description supports Markdown formatting, which is helpful for documenting what the question shows and why it matters.
</Info>

## Question metadata and information

Click the **info** icon in the upper right to view details about your question:

### Overview tab

* **Description**: Add context about what the question shows (supports Markdown)
* **Created by**: Who originally created the question
* **Last edited by**: Who made the most recent changes
* **Saved in**: The collection or dashboard containing the question
* **Based on**: The underlying database and table
* **Entity ID**: Unique identifier used for serialization across Metabase instances

### History tab

Metabase automatically tracks the last 15 versions of each question. You can:

* View what changed in each version
* See who made the changes and when
* Revert to a previous version if needed

## Working with questions

<AccordionGroup>
  <Accordion title="Preview results while building">
    As you build your question, click the **Preview** button (play icon) next to each step to see the first 10 rows of results. This helps you validate your query before running the full query.
  </Accordion>

  <Accordion title="Explore with drill-through">
    Click on data points in your visualizations to drill deeper:

    * See underlying records
    * Filter by clicked values
    * Break out by different dimensions
    * View related data

    Drill-through options vary based on your chart type and data.
  </Accordion>

  <Accordion title="Convert between query builder and SQL">
    You can view the SQL that Metabase generates for query builder questions by clicking **View SQL** in the top right. You can also convert a query builder question to SQL (this is one-way - you can't convert SQL back to query builder).
  </Accordion>

  <Accordion title="Turn questions into models">
    Questions saved to collections can be promoted to models to signal they're good starting points for analysis. Models can have additional metadata and serve as curated data sources.
  </Accordion>
</AccordionGroup>

## Question permissions

<Warning>
  Your ability to create, edit, and save questions depends on your permissions:

  * **View access**: See the question and its results
  * **Curate access**: Edit and move questions in a collection
  * **Data access**: Query specific databases and tables
</Warning>

To save questions to a collection, you need curate access. To create questions using the SQL editor, you need both query builder and native query permissions for that database.

## Checking for breaking changes

<Info>
  This feature is available on paid plans.
</Info>

When saving changes to a question, Metabase checks if your changes would break dependent entities. For example:

* Removing a column that other questions reference
* Renaming fields used in downstream questions
* Changing data types in ways that break filters

Metabase will warn you before saving changes that could break other questions, dashboards, or metrics.

## Moving questions between collections and dashboards

You can move a question from a collection to a dashboard if:

* No other dashboards use that question, OR
* You have curate access to all dashboards using that question

When moving a question from a collection to a dashboard, Metabase will warn you if the question is used elsewhere and ask you to confirm the removal.

## Best practices

<CardGroup cols={2}>
  <Card title="Name questions clearly" icon="tag">
    Use descriptive names that explain what the question shows. "Monthly revenue by region" is better than "Revenue question".
  </Card>

  <Card title="Add descriptions" icon="file-lines">
    Document the purpose, methodology, and any important caveats. Future you (and your teammates) will thank you.
  </Card>

  <Card title="Start simple" icon="seedling">
    Build questions incrementally. Start with a simple query, verify the results, then add complexity.
  </Card>

  <Card title="Preview as you build" icon="eye">
    Use the preview button to check results after each step. Catch issues early before running expensive queries.
  </Card>
</CardGroup>

## Next steps

<CardGroup cols={2}>
  <Card title="Query builder" icon="wand-magic-sparkles" href="/questions/query-builder">
    Learn how to use the graphical query builder to create questions without SQL
  </Card>

  <Card title="SQL editor" icon="code" href="/questions/sql-editor">
    Explore the native SQL editor for advanced queries and complex analysis
  </Card>

  <Card title="Creating dashboards" icon="table-columns" href="/dashboards/creating-dashboards">
    Organize your questions into dashboards for monitoring and reporting
  </Card>

  <Card title="Alerts" icon="bell" href="/dashboards/alerts">
    Set up alerts to get notified when your data changes
  </Card>
</CardGroup>