> ## Documentation Index
> Fetch the complete documentation index at: https://resources.athenaintel.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Create a semantic model from lakehouse tables
> Select tables in a shared namespace and create a named semantic model
You create a semantic model from lakehouse tables when you want to describe their data as dimensions and measures for analysis. Dimensions describe values you group or filter by. Measures describe calculations such as counts or sums. Athena generates an initial model from the tables you select, so you can review the definitions instead of starting with an empty model.
## Choose your source tables
You need edit access to the namespace to use its creation action. Your deployment also needs lakehouse support configured for semantic models.
1. Open the lakehouse namespace asset that contains your tables.
2. Review the tables and their columns. Choose the tables you need for your analysis.
3. Select **Create semantic model** in the namespace header.
4. Select the checkboxes beside the tables you want to include. **All tables** selects the whole list.
5. The **Name** field starts with a default based on the namespace, such as *sales model*. Keep it, type a different name, or clear it to use the default.
6. Select **Create**. Athena opens the new semantic model after creation succeeds.
1. Open the lakehouse namespace
2. Select Create semantic model
3. Check the tables to include
4. Keep or change the Name
5. Select Create
6. Review the generated model
## Review the generated model
1. Review the generated table definitions in the model. They point to the selected lakehouse tables.
2. Check the dimensions against the source columns. Review any generated primary key rather than treating it as proof that the source values are unique.
3. Review the generated measures. The initial model includes a row count and generates sum and average measures for numeric columns whose names end in `_amount`, `_price`, `_cost`, `_total`, `_value`, `_revenue`, `_fee`, `_qty`, `_quantity` or `_count`. Those columns become measures, not dimensions.
4. Refine the business names and calculations before using the model for analysis.
5. The new model is a draft. Select **Deploy** to make its measures queryable. **Browser**, **Diagram** and **Query** stay unavailable until you deploy.
The generated definitions are a starting point. Creating the asset does not confirm that every generated calculation matches your business rules.
## Limits
* Select at least one table and at most 100 tables from the namespace.
* If the table list is unavailable, check your namespace access. If the namespace has no tables, land data there before creating the model.
* Creation checks your access to the namespace and each selected table in your current workspace.
* Later queries and deployments recheck the lakehouse source bindings. Sharing the model does not replace access to its source namespace.
* If creation fails, the dialog shows the error. Resolve the access or deployment issue before trying again.
## Related links
* [SQL queries](./query)
* [Spaces and assets](../concepts/spaces-and-assets)
* [Databricks connection](../../integrations/databricks)
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.