Skip to content

Demo from Snowflake

These scenarios are based upon several examples of the use of the semantic hub and semantic model catalog capabilities in the product.

One example is the ability for the intelligent feature of the semantic model catalog to import a model model with related metadata which is derived from the database itself (e.g., Snowflake) and:

  • Include various tables, columns, metrics and joins from the source database model.

Self Service Snowflake Reverse Engineering

Overview

The demo is based on the Snowflake database and schema: SEMANTIC_HUB.MINI_RETAIL_DW.

Importing the Schema into a Semantic Model Catalog

Steps

  1. Import the model from Snowflake
  2. Open the semantic model catalog named Snowflake Based Semantic Model:
  3. Click Data Warehouse
  4. Database Type: Snowflake
  5. Model: Snowflake Database (via JDBC)
  6. Catalog: SEMANTIC_HUB
  7. Schema: MINI_RETAIL_DW

You must already have a model imported from Tableau from which to pick.

  1. Click OK

  2. Once completed go to the Editor tab.

Example

Create and Populate the Semantic Model

We already have the Snowflake Model imported, providing the semantic model catalog with connection information and browsing of tables.

Open the semantic model catalog named Snowflake Based Semantic Model:

Click Data Warehouse:

Specify as above.

Click OK.

Then select all of the tables and click OK.

After the import and merge finishes you will have:

Go to the Editor tab.

Right-click on the Tableau Based Semantic Model in the root of the Catalog and select Open Diagram and click the Fact Sales Txn table:

Click START EDITING.

Update properties for this model

Then go to the Connection at the top of the Catalog panel:

Scroll down to the Database Deployment Parameters and enter the following:

  • Deployment Database: SEMANTIC_HUB
  • Deployment Schema: SEMANTIC_VIEWS_TESTING

Under BI Deployment click the plus sign (ADD) and pick Tableau and enter the following (from the pick list and browse):

  • Tableau Server (Repository) and Cloud
  • Demo Semantic Hub John

Click SAVE.

Correct Any Issues

Note, there is one warning/error in this model reverse engineered from the Snowflake database:

You can see the join loop, which is not supported at this time.

Simply select it in the diagram, right-click and select Delete

and click SAVE:

So, the model is without issues.

Create a Subset Model

You may create as many models to deploy as you want. In this case, we will build one around the same central fact as with the Tableau and Power BI examples.

Right-click on Model at the bottom of the Catalog panel and select New Model:

and name it "FactSalesDeploymentModel" and double-click on it to open it in another tab:

Pick the Snowflake Database (via JDBC) as the Connection. Click SAVE.

Then, expand the model in the tree, right-click on the Fact Tables and select New Fact Table and in the Properties panel use the pull-down to select the Fact Sales Txn model:

Click SAVE and STOP EDITING.

Deployment

You may deploy the model to a database, such as Snowflake, or deploy to a BI tool, such as Tableau. See the example of an imported model from Tableau and deployment.