# How to query tables from multiple databases in Snowflake?

> Use a Query Model in Holistics to reference tables in a second Snowflake database, since each data source can only connect to one database by default.

## Question

I'm using Snowflake with Holistics. Snowflake supports [joining tables from 2 different databases](https://community.snowflake.com/s/question/0D50Z00007AptG4SAJ/is-there-anyway-to-achieve-cross-db-join-in-snowflake) in the same cloud instance. But in Holistics, each data source has to link to just one Snowflake database.

How can I run join queries on tables that sit in different Snowflake databases?

## Context

Currently when connecting to Snowflake, Holistics requires you to specify one database name. While this works well for a default database, it currently doesn't support selecting multiple databases.

Because of this, when creating data models (table models) in Holistics, you can only choose tables in database A but not database B. All the tables in database B are not visible to the end users.

Because of this, you don't know how to pull data from database B for Holistics modeling.

## Solution

### Create database views in db1 that refer to tables in db2

This is a workaround that we can implement at the moment. Instead of creating a 'table model' to link to a database table, you can create a [Query Model](/docs/query-models.md)

```aml
Model users_db2 {
  type: 'query'
  data_source_name: 'mysnowflake'

  dimension field1 {
    type: 'number'
  }
  ...

  query: @sql
		select * from db2.public.users;
  ;;
}
```

In the above model, it queries data from database B `db2.some_table` instead of the current database A.

Once created, just treat the model as per a normal table model and create relationships to link them with other models in database A.

### The Proper Solution

*(This has not been supported. Please reach out to Support for more info)*

The proper solution is to fix the above "Add Model" dialog, so that when adding a table model, you can select tables in other databases.

This way you can bring tables from two different databases into a single modeling layer. These models then can be combined (relationships) or joined together.