In this post we will discuss using dbt with the Cloudera Data Platform, and show you how to get started by connecting dbt to your Impala Data Warehouse. You’ll also find links to a dbt example project that you can use to bootstrap your dbt journey.
Cloudera Data Warehouse (CDW) is a CDP Public Cloud service for self-service creation of independent data warehouses and data marts that autoscale up and down to meet your varying workload demands. The Data Warehouse service provides isolated compute instances for each data warehouse/mart, automatic optimization, and enables you to save costs while meeting SLAs. Both Apache Impala and Apache Hive are available through Cloudera Data Warehouse.
dbt is quickly gaining popularity as a key component of the modern data stack; a tool that enables the creation of data pipelines & analytics projects using only SQL.
In the words of dbtLabs:
“dbt™ is a transformation workflow that lets teams quickly and collaboratively deploy analytics code following software engineering best practices like modularity, portability, CI/CD, and documentation. Now anyone who knows SQL can build production-grade data pipelines.”
dbt leverages your existing warehouse to run your workflows, meaning you avoid the complexities of additional hardware/tools/clusters for extracting, transforming and then loading back into the warehouse.
To use dbt with Impala, you need the following python packages: dbt-core, dbt-impala and impyla.
Start by cloning the demo repository at https://github.com/cloudera/dbt-impala-example with:
git clone https://github.com/cloudera/dbt-impala-example.git
Inside this repo is a file called requirements.txt, which you can use to install the python dependencies. Install the requirements using pip:
pip install -r requirements.txt
To work with Impala in CDP, we need two things - an Impala Virtual Warehouse (part of Cloudera Data Warehouse) and a user to run queries.
For this demo, we will use a Machine User created inside CDP as the user running queries. Inside CDP > User Management, add a new Machine User and set the workload password.
The steps for this are documented here:
With the user created & the workload password set, take a note of the Workload username & password. Notice in the below screenshot, for a Machine User called ‘cia_test_user’ the workload username is ‘srv_cia_test_user’.
Keep the workload user & password details handy for later.
We will be using Impala through Cloudera Data Warehouse - a cloud-native, auto-scaling deployment of Impala.
Start by activating your CDW Environment as documented here: https://docs.cloudera.com/data-warehouse/cloud/aws-environments/topics/dw-activating-environments-4-...
This will create a default database catalog, which we will use in this demo. You are able to create non-default database catalogs, as documented here: https://docs.cloudera.com/data-warehouse/cloud/managing-warehouses/topics/dw-adding-new-database-cat...
Next, create an Impala Virtual Warehouse connected to the default database catalog, as documented here: https://docs.cloudera.com/data-warehouse/cloud/managing-warehouses/topics/dw-adding-new-virtual-ware...
The following settings we will be used for Impala Virtual Warehouse in this demo:
Once created, you should see that the Virtual Warehouse enters the running state
From here, select the 3 dots, and then Copy JDBC URL. In my case, this looks like:
Keep this URL handy for later.
dbt requires that we configure a profile that defines how to connect to our data warehouse. For this, we need the workload credentials & Impala connection details we collected earlier.
The profile lives in a `.dbt` directory in your home directory and is called `profiles.yml`. On Linux, this would look like `~/.dbt/profiles.yml`. If you haven't used dbt before, create the directory with `mkdir ~/.dbt` and create the `profiles.yml` file with your favourite text editor.
You can learn more about the dbt profile from the dbt docs here https://docs.getdbt.com/dbt-cli/configure-your-profile
Use the following template for the contents of the file:
dbt_impala_demo: outputs: dev: type: impala host: <Impala Hostname> port: 443 dbname: dbt_demo schema: dbt_demo user: <Workload Username> password: <Workload Password> auth_type: ldap use_http_transport: true use_ssl: true http_path: cliservice target: dev
First, add your Workload user/pass to the “user” and “password” fields.
Next, we need to extract the Impala hostname from the JDBC URL we copied earlier. We do not want to use the entire JDBC URL.
Given the above JDBC URL - we want to extract the hostname between the protocol (“jdbc:impala://”) and the port (“:443”). The result is:
Use the extracted hostname for the “host” field in the template.
My completed profile looks like this:
dbt_impala_demo: outputs: dev: type: impala host: coordinator-cia-dbt-impala.dw-ciadev.a465-9q4k.cloudera.site port: 443 dbname: dbt_demo schema: dbt_demo user: srv_cia_test_user password: Password123! auth_type: ldap use_http_transport: true use_ssl: true http_path: cliservice target: dev
To ensure we’ve configured our profile correctly, let’s run a connection test. For this we use the command:
In the output of this command, you should see the following:
Connection: host: coordinator-cia-dbt-impala.dw-ciadev.a465-9q4k.cloudera.site port: 443 database: dbt_demo schema: dbt_demo username: srv_cia_test_user Connection test: [OK connection ok]
This confirms a successful connection to the Impala warehouse.
In the example repo we cloned at the start, we have a demo dbt project called ‘dbt_impala_demo’.
Inside this demo project, we can issue dbt commands to run parts of the project. The demo project contains examples for: generating fake data, tests, seeds, sources, view models & incremental table models.
To run the seeds, use:
To run the tests, use:
To run the models, use:
You can also generate dbt documentation using:
dbt docs generate
The README for the example repo going into further detail about using the demo project: https://github.com/cloudera/dbt-impala-example/blob/main/README.md and we’ll cover it in further detail in a later Community post.
We have covered a quick intro to dbt, and worked through setting up our environment to get dbt connected to Cloudera Data Warehouse. We’ve also introduced the example repo to help bootstrap your journey to CDP.
In a later post we’ll cover the example repo in more detail and demonstrate some real use cases for dbt.
If you have any questions or feedback related to dbt on the Cloudera Data Platform, please reach out to us via this community, or drop us an email at email@example.com
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.