Quickstart for dbt Cloud and Azure Synapse Analytics
- 1 Introduction
- 2 Load data into your Azure Synapse Analytics
- 3 Connect dbt Cloud to Azure Synapse Analytics
- 4 Set up a dbt Cloud managed repository
- 5 Initialize your dbt project and start developing
- 6 Build your first model
- 7 Change the way your model is materialized
- 8 Delete the example models
- 9 Build models on top of other models
- 10 Add tests to your models
- 11 Document your models
- 12 Commit your changes
- 13 Deploy dbt
Introduction
In this quickstart guide, you'll learn how to use dbt Cloud with Azure Synapse Analytics. It will show you how to:
- Load the Jaffle Shop sample data (provided by dbt Labs) into your Azure Synapse Analytics warehouse.
- Connect dbt Cloud to Azure Synapse Analytics.
- Turn a sample query into a model in your dbt project. A model in dbt is a SELECT statement.
- Add tests to your models.
- Document your models.
- Schedule a job to run.
Prerequisites
- You have a dbt Cloud account.
- You have an Azure Synapse Analytics account. For a free trial, refer to Synapse Analytics in the Microsoft docs.
- As a Microsoft admin, you’ve enabled service principal authentication. You must add the service principal to the Synapse workspace with either a Member (recommended) or Admin permission set. For details, refer to Create a service principal using the Azure portal in the Microsoft docs. dbt Cloud needs these authentication credentials to connect to Azure Synapse Analytics.
Related content
Load data into your Azure Synapse Analytics
-
Log in to your Azure portal account.
-
On the home page, select the SQL databases tile.
-
From the SQL databases page, navigate to your organization’s workspace or create a new workspace; refer to Create a Synapse workspace in the Microsoft docs for more details.
-
From the workspace's sidebar, select Data. Click the three dot menu on your database and select New SQL script to open the SQL editor.
-
Copy these statements into the SQL editor to load the Jaffle Shop example data:
CREATE TABLE dbo.customers
(
[ID] [bigint],
[FIRST_NAME] [varchar](8000),
[LAST_NAME] [varchar](8000)
);
COPY INTO [dbo].[customers]
FROM 'https://dbtlabsynapsedatalake.blob.core.windows.net/dbt-quickstart-public/jaffle_shop_customers.parquet'
WITH (
FILE_TYPE = 'PARQUET'
);
CREATE TABLE dbo.orders
(
[ID] [bigint],
[USER_ID] [bigint],
[ORDER_DATE] [date],
[STATUS] [varchar](8000)
);
COPY INTO [dbo].[orders]
FROM 'https://dbtlabsynapsedatalake.blob.core.windows.net/dbt-quickstart-public/jaffle_shop_orders.parquet'
WITH (
FILE_TYPE = 'PARQUET'
);
CREATE TABLE dbo.payments
(
[ID] [bigint],
[ORDERID] [bigint],
[PAYMENTMETHOD] [varchar](8000),
[STATUS] [varchar](8000),
[AMOUNT] [bigint],
[CREATED] [date]
);
COPY INTO [dbo].[payments]
FROM 'https://dbtlabsynapsedatalake.blob.core.windows.net/dbt-quickstart-public/stripe_payments.parquet'
WITH (
FILE_TYPE = 'PARQUET'
);
Connect dbt Cloud to Azure Synapse Analytics
- Create a new project in dbt Cloud. Open the gear menu in the top right corner, select Account settings and click + New Project.
- Enter a project name and click Continue.
- Choose Synapse as your connection and click Next.
- In the Configure your environment section, enter the Settings for your new project:
- Server — Use the service principal's Synapse host name value (without the trailing
, 1433
string) for the Synapse test endpoint. - Port — 1433 (which is the default).
- Database — Use the service principal's database value for the Synapse test endpoint.
- Server — Use the service principal's Synapse host name value (without the trailing
- Enter the Development credentials for your new project:
- Authentication — Choose Service Principal from the dropdown.
- Tenant ID — Use the service principal’s Directory (tenant) id as the value.
- Client ID — Use the service principal’s application (client) ID id as the value.
- Client secret — Use the service principal’s client secret (not the client secret id) as the value.
- Click Test connection. This verifies that dbt Cloud can access your Azure Synapse Analytics account.
- Click Next when the test succeeds. If it failed, you might need to check your Microsoft service principal.
Set up a dbt Cloud managed repository
When you develop in dbt Cloud, you can leverage Git to version control your code.
To connect to a repository, you can either set up a dbt Cloud-hosted managed repository or directly connect to a supported git provider. Managed repositories are a great way to trial dbt without needing to create a new repository. In the long run, it's better to connect to a supported git provider to use features like automation and continuous integration.
To set up a managed repository:
- Under "Setup a repository", select Managed.
- Type a name for your repo such as
bbaggins-dbt-quickstart
- Click Create. It will take a few seconds for your repository to be created and imported.
- Once you see the "Successfully imported repository," click Continue.
Initialize your dbt project and start developing
Now that you have a repository configured, you can initialize your project and start development in dbt Cloud:
- Click Start developing in the IDE. It might take a few minutes for your project to spin up for the first time as it establishes your git connection, clones your repo, and tests the connection to the warehouse.
- Above the file tree to the left, click Initialize dbt project. This builds out your folder structure with example models.
- Make your initial commit by clicking Commit and sync. Use the commit message
initial commit
and click Commit Changes. This creates the first commit to your managed repo and allows you to open a branch where you can add new dbt code. - You can now directly query data from your warehouse and execute
dbt run
. You can try this out now:- In the command line bar at the bottom, enter
dbt run
and click Enter. You should see adbt run succeeded
message.
- In the command line bar at the bottom, enter
Build your first model
-
Under Version Control on the left, click Create branch. You can name it
add-customers-model
. You need to create a new branch since the main branch is set to read-only mode. -
Click the three dot menu (...) next to the
models
directory, then select Create file. -
Name the file
customers.sql
, then click Create. -
Copy the following query into the file and click Save.
customers.sqlwith customers as (
select
ID as customer_id,
FIRST_NAME as first_name,
LAST_NAME as last_name
from dbo.customers
),
orders as (
select
ID as order_id,
USER_ID as customer_id,
ORDER_DATE as order_date,
STATUS as status
from dbo.orders
),
customer_orders as (
select
customer_id,
min(order_date) as first_order_date,
max(order_date) as most_recent_order_date,
count(order_id) as number_of_orders
from orders
group by customer_id
),
final as (
select
customers.customer_id,
customers.first_name,
customers.last_name,
customer_orders.first_order_date,
customer_orders.most_recent_order_date,
coalesce(customer_orders.number_of_orders, 0) as number_of_orders
from customers
left join customer_orders on customers.customer_id = customer_orders.customer_id
)
select * from final -
Enter
dbt run
in the command prompt at the bottom of the screen. You should get a successful run and see the three models.
Later, you can connect your business intelligence (BI) tools to these views and tables so they only read cleaned up data rather than raw data in your BI tool.
FAQs
To check out the SQL that dbt is running, you can look in:
- dbt Cloud:
- Within the run output, click on a model name, and then select "Details"
- dbt Core:
- The
target/compiled/
directory for compiledselect
statements - The
target/run/
directory for compiledcreate
statements - The
logs/dbt.log
file for verbose logging.
- The
By default, dbt builds models in your target schema. To change your target schema:
- If you're developing in dbt Cloud, these are set for each user when you first use a development environment.
- If you're developing with dbt Core, this is the
schema:
parameter in yourprofiles.yml
file.
If you wish to split your models across multiple schemas, check out the docs on using custom schemas.
Note: on BigQuery, dataset
is used interchangeably with schema
.
Nope! dbt will check if the schema exists when it runs. If the schema does not exist, dbt will create it for you.
Nope! The SQL that dbt generates behind the scenes ensures that any relations are replaced atomically (i.e. your business users won't experience any downtime).
The implementation of this varies on each warehouse, check out the logs to see the SQL dbt is executing.
If there's a mistake in your SQL, dbt will return the error that your database returns.
$ dbt run --select customers
Running with dbt=0.15.0
Found 3 models, 9 tests, 0 snapshots, 0 analyses, 133 macros, 0 operations, 0 seed files, 0 sources
14:04:12 | Concurrency: 1 threads (target='dev')
14:04:12 |
14:04:12 | 1 of 1 START view model dbt_alice.customers.......................... [RUN]
14:04:13 | 1 of 1 ERROR creating view model dbt_alice.customers................. [ERROR in 0.81s]
14:04:13 |
14:04:13 | Finished running 1 view model in 1.68s.
Completed with 1 error and 0 warnings:
Database Error in model customers (models/customers.sql)
Syntax error: Expected ")" but got identifier `your-info-12345` at [13:15]
compiled SQL at target/run/jaffle_shop/customers.sql
Done. PASS=0 WARN=0 ERROR=1 SKIP=0 TOTAL=1
Any models downstream of this model will also be skipped. Use the error message and the compiled SQL to debug any errors.
Change the way your model is materialized
One of the most powerful features of dbt is that you can change the way a model is materialized in your warehouse, simply by changing a configuration value. You can change things between tables and views by changing a keyword rather than writing the data definition language (DDL) to do this behind the scenes.
By default, everything gets created as a view. You can override that at the directory level so everything in that directory will materialize to a different materialization.
-
Edit your
dbt_project.yml
file.-
Update your project
name
to:dbt_project.ymlname: 'jaffle_shop'
-
Configure
jaffle_shop
so everything in it will be materialized as a table; and configureexample
so everything in it will be materialized as a view. Update yourmodels
config block to:dbt_project.ymlmodels:
jaffle_shop:
+materialized: table
example:
+materialized: view -
Click Save.
-
-
Enter the
dbt run
command. Yourcustomers
model should now be built as a table!infoTo do this, dbt had to first run a
drop view
statement (or API call on BigQuery), then acreate table as
statement. -
Edit
models/customers.sql
to override thedbt_project.yml
for thecustomers
model only by adding the following snippet to the top, and click Save:models/customers.sql{{
config(
materialized='view'
)
}}
with customers as (
select
id as customer_id
...
) -
Enter the
dbt run
command. Your model,customers
, should now build as a view.- BigQuery users need to run
dbt run --full-refresh
instead ofdbt run
to full apply materialization changes.
- BigQuery users need to run
-
Enter the
dbt run --full-refresh
command for this to take effect in your warehouse.
FAQs
dbt ships with five materializationsThe exact Data Definition Language (DDL) that dbt will use when creating the model’s equivalent in a data warehouse.: view
, table
, incremental
, ephemeral
and materialized_view
.
Check out the documentation on materializations for more information on each of these options.
You can also create your own custom materializations, if required however this is an advanced feature of dbt.
Start out with viewsA view (as opposed to a table) is a defined passthrough SQL query that can be run against a database (or data warehouse)., and then change models to tables when required for performance reasons (i.e. downstream queries have slowed).
Check out the docs on materializations for advice on when to use each materializationThe exact Data Definition Language (DDL) that dbt will use when creating the model’s equivalent in a data warehouse..
You can also configure:
- tags to support easy categorization and graph selection
- custom schemas to split your models across multiple schemas
- aliases if your viewA view (as opposed to a table) is a defined passthrough SQL query that can be run against a database (or data warehouse)./tableIn simplest terms, a table is the direct storage of data in rows and columns. Think excel sheet with raw values in each of the cells. name should differ from the filename
- Snippets of SQL to run at the start or end of a model, known as hooks
- Warehouse-specific configurations for performance (e.g.
sort
anddist
keys on Redshift,partitions
on BigQuery)
Check out the docs on model configurations to learn more.
Delete the example models
You can now delete the files that dbt created when you initialized the project:
-
Delete the
models/example/
directory. -
Delete the
example:
key from yourdbt_project.yml
file, and any configurations that are listed under it.dbt_project.yml# before
models:
jaffle_shop:
+materialized: table
example:
+materialized: viewdbt_project.yml# after
models:
jaffle_shop:
+materialized: table -
Save your changes.
FAQs
If you delete a model from your dbt project, dbt does not automatically drop the relation from your schema. This means that you can end up with extra objects in schemas that dbt creates, which can be confusing to other users.
(This can also happen when you switch a model from being a viewA view (as opposed to a table) is a defined passthrough SQL query that can be run against a database (or data warehouse). or tableIn simplest terms, a table is the direct storage of data in rows and columns. Think excel sheet with raw values in each of the cells., to ephemeral)
When you remove models from your dbt project, you should manually drop the related relations from your schema.
You might have forgotten to nest your configurations under your project name, or you might be trying to apply configurations to a directory that doesn't exist.
Check out this article to understand more.
Build models on top of other models
As a best practice in SQL, you should separate logic that cleans up your data from logic that transforms your data. You have already started doing this in the existing query by using common table expressions (CTEs).
Now you can experiment by separating the logic out into separate models and using the ref function to build models on top of other models:
-
Create a new SQL file,
models/stg_customers.sql
, with the SQL from thecustomers
CTE in our original query. -
Create a second new SQL file,
models/stg_orders.sql
, with the SQL from theorders
CTE in our original query.models/stg_customers.sqlselect
ID as customer_id,
FIRST_NAME as first_name,
LAST_NAME as last_name
from dbo.customersmodels/stg_orders.sqlselect
ID as order_id,
USER_ID as customer_id,
ORDER_DATE as order_date,
STATUS as status
from dbo.orders -
Edit the SQL in your
models/customers.sql
file as follows:models/customers.sqlwith customers as (
select * from {{ ref('stg_customers') }}
),
orders as (
select * from {{ ref('stg_orders') }}
),
customer_orders as (
select
customer_id,
min(order_date) as first_order_date,
max(order_date) as most_recent_order_date,
count(order_id) as number_of_orders
from orders
group by customer_id
),
final as (
select
customers.customer_id,
customers.first_name,
customers.last_name,
customer_orders.first_order_date,
customer_orders.most_recent_order_date,
coalesce(customer_orders.number_of_orders, 0) as number_of_orders
from customers
left join customer_orders on customers.customer_id = customer_orders.customer_id
)
select * from final -
Execute
dbt run
.This time, when you performed a
dbt run
, separate views/tables were created forstg_customers
,stg_orders
andcustomers
. dbt inferred the order to run these models. Becausecustomers
depends onstg_customers
andstg_orders
, dbt buildscustomers
last. You do not need to explicitly define these dependencies.
FAQs
To run one model, use the --select
flag (or -s
flag), followed by the name of the model:
$ dbt run --select customers
Check out the model selection syntax documentation for more operators and examples.
Within one project: yes! To build dependencies between resources (such as models, seeds, and snapshots), you need to use the ref
function, and pass in the resource name as an argument. dbt uses that resource name to uniquely resolve the ref
to a specific resource. As a result, these resource names need to be unique, even if they are in distinct folders.
A resource in one project can have the same name as a resource in another project (installed as a dependency). dbt uses the project name to uniquely identify each resource. We call this "namespacing." If you ref
a resource with a duplicated name, it will resolve to the resource within the same namespace (package or project), or raise an error because of an ambiguous reference. Use two-argument ref
to disambiguate references by specifying the namespace.
Those resource will still need to land in distinct locations in the data warehouse. Read the docs on custom aliases and custom schemas for details on how to achieve this.
There's no one best way to structure a project! Every organization is unique.
If you're just getting started, check out how we (dbt Labs) structure our dbt projects.
Add tests to your models
Adding tests to a project helps validate that your models are working correctly.
To add tests to your project:
-
Create a new YAML file in the
models
directory, namedmodels/schema.yml
-
Add the following contents to the file:
models/schema.ymlversion: 2
models:
- name: customers
columns:
- name: customer_id
tests:
- unique
- not_null
- name: stg_customers
columns:
- name: customer_id
tests:
- unique
- not_null
- name: stg_orders
columns:
- name: order_id
tests:
- unique
- not_null
- name: status
tests:
- accepted_values:
values: ['placed', 'shipped', 'completed', 'return_pending', 'returned']
- name: customer_id
tests:
- not_null
- relationships:
to: ref('stg_customers')
field: customer_id -
Run
dbt test
, and confirm that all your tests passed.
When you run dbt test
, dbt iterates through your YAML files, and constructs a query for each test. Each query will return the number of records that fail the test. If this number is 0, then the test is successful.
FAQs
Out of the box, dbt ships with the following tests:
unique
not_null
accepted_values
relationships
(i.e. referential integrity)
You can also write your own custom schema data tests.
Some additional custom schema tests have been open-sourced in the dbt-utils package, check out the docs on packages to learn how to make these tests available in your project.
Note that although you can't document data tests as of yet, we recommend checking out this dbt Core discussion where the dbt community shares ideas.
Running tests on one model looks very similar to running a model: use the --select
flag (or -s
flag), followed by the name of the model:
dbt test --select customers
Check out the model selection syntax documentation for full syntax, and test selection examples in particular.
To debug a failing test, find the SQL that dbt ran by:
-
dbt Cloud:
- Within the test output, click on the failed test, and then select "Details"
-
dbt Core:
- Open the file path returned as part of the error message.
- Navigate to the
target/compiled/schema_tests
directory for all compiled test queries
Copy the SQL into a query editor (in dbt Cloud, you can paste it into a new Statement
), and run the query to find the records that failed.
No! You can name this file whatever you want (including whatever_you_want.yml
), so long as:
- The file is in your
models/
directory¹ - The file has
.yml
extension
Check out the docs for more information.
¹If you're declaring properties for seeds, snapshots, or macros, you can also place this file in the related directory — seeds/
, snapshots/
and macros/
respectively.
Once upon a time, the structure of these .yml
files was very different (s/o to anyone who was using dbt back then!). Adding version: 2
allowed us to make this structure more extensible.
Resource yml files do not currently require this config. We only support version: 2
if it's specified. Although we do not expect to update yml files to version: 3
soon, having this config will make it easier for us to introduce new structures in the future
We recommend that every model has a test on a primary keyA primary key is a non-null column in a database object that uniquely identifies each row., that is, a column that is unique
and not_null
.
We also recommend that you test any assumptions on your source data. For example, if you believe that your payments can only be one of three payment methods, you should test that assumption regularly — a new payment method may introduce logic errors in your SQL.
In advanced dbt projects, we recommend using sources and running these source data-integrity tests against the sources rather than models.
You should run your tests whenever you are writing new code (to ensure you haven't broken any existing models by changing SQL), and whenever you run your transformations in production (to ensure that your assumptions about your source data are still valid).
Document your models
Adding documentation to your project allows you to describe your models in rich detail, and share that information with your team. Here, we're going to add some basic documentation to our project.
-
Update your
models/schema.yml
file to include some descriptions, such as those below.models/schema.ymlversion: 2
models:
- name: customers
description: One record per customer
columns:
- name: customer_id
description: Primary key
tests:
- unique
- not_null
- name: first_order_date
description: NULL when a customer has not yet placed an order.
- name: stg_customers
description: This model cleans up customer data
columns:
- name: customer_id
description: Primary key
tests:
- unique
- not_null
- name: stg_orders
description: This model cleans up order data
columns:
- name: order_id
description: Primary key
tests:
- unique
- not_null
- name: status
tests:
- accepted_values:
values: ['placed', 'shipped', 'completed', 'return_pending', 'returned']
- name: customer_id
tests:
- not_null
- relationships:
to: ref('stg_customers')
field: customer_id -
Run
dbt docs generate
to generate the documentation for your project. dbt introspects your project and your warehouse to generate a JSONJSON (JavaScript Object Notation) is a minimal format for semi-structured data used to capture relationships between fields and values. file with rich documentation about your project.
- Click the book icon in the Develop interface to launch documentation in a new tab.
FAQs
If you need more than a sentence to explain a model, you can:
- Split your description over multiple lines using
>
. Interior line breaks are removed and Markdown can be used. This method is recommended for simple, single-paragraph descriptions:
version: 2
models:
- name: customers
description: >
Lorem ipsum **dolor** sit amet, consectetur adipisicing elit, sed do eiusmod
tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,
quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo
consequat.
- Split your description over multiple lines using
|
. Interior line breaks are maintained and Markdown can be used. This method is recommended for more complex descriptions:
version: 2
models:
- name: customers
description: |
### Lorem ipsum
* dolor sit amet, consectetur adipisicing elit, sed do eiusmod
* tempor incididunt ut labore et dolore magna aliqua.
- Use a docs block to write the description in a separate Markdown file. b
If you're using dbt Cloud to deploy your project and have the Team or Enterprise plan, you can use dbt Explorer to view your project's resources (such as models, tests, and metrics) and their lineageData lineage provides a holistic view of how data moves through an organization, where it’s transformed and consumed. to gain a better understanding of its latest production state.
Access dbt Explorer in dbt Cloud by clicking the Explore link in the navigation. You can have up to 5 read-only users access the documentation for your project.
dbt Cloud developer plan and dbt Core users can use dbt Docs, which generates basic documentation but it doesn't offer the same speed, metadata, or visibility as dbt Explorer.
Commit your changes
Now that you've built your customer model, you need to commit the changes you made to the project so that the repository has your latest code.
If you edited directly in the protected primary branch:
- Click the Commit and sync git button. This action prepares your changes for commit.
- A modal titled Commit to a new branch will appear.
- In the modal window, name your new branch
add-customers-model
. This branches off from your primary branch with your new changes. - Add a commit message, such as "Add customers model, tests, docs" and and commit your changes.
- Click Merge this branch to main to add these changes to the main branch on your repo.
If you created a new branch before editing:
- Since you already branched out of the primary protected branch, go to Version Control on the left.
- Click Commit and sync to add a message.
- Add a commit message, such as "Add customers model, tests, docs."
- Click Merge this branch to main to add these changes to the main branch on your repo.
Deploy dbt
Use dbt Cloud's Scheduler to deploy your production jobs confidently and build observability into your processes. You'll learn to create a deployment environment and run a job in the following steps.
Create a deployment environment
- In the upper left, select Deploy, then click Environments.
- Click Create Environment.
- In the Name field, write the name of your deployment environment. For example, "Production."
- In the dbt Version field, select the latest version from the dropdown.
- Under Deployment connection, enter the name of the dataset you want to use as the target, such as "Analytics". This will allow dbt to build and work with that dataset. For some data warehouses, the target dataset may be referred to as a "schema".
- Click Save.
Create and run a job
Jobs are a set of dbt commands that you want to run on a schedule. For example, dbt build
.
As the jaffle_shop
business gains more customers, and those customers create more orders, you will see more records added to your source data. Because you materialized the customers
model as a table, you'll need to periodically rebuild your table to ensure that the data stays up-to-date. This update will happen when you run a job.
- After creating your deployment environment, you should be directed to the page for a new environment. If not, select Deploy in the upper left, then click Jobs.
- Click Create one and provide a name, for example, "Production run", and link to the Environment you just created.
- Scroll down to the Execution Settings section.
- Under Commands, add this command as part of your job if you don't see it:
dbt build
- Select the Generate docs on run checkbox to automatically generate updated project docs each time your job runs.
- For this exercise, do not set a schedule for your project to run — while your organization's project should run regularly, there's no need to run this example project on a schedule. Scheduling a job is sometimes referred to as deploying a project.
- Select Save, then click Run now to run your job.
- Click the run and watch its progress under "Run history."
- Once the run is complete, click View Documentation to see the docs for your project.
Congratulations 🎉! You've just deployed your first dbt project!
FAQs
If you're using dbt Cloud, we recommend setting up email and Slack notifications (Account Settings > Notifications
) for any failed runs. Then, debug these runs the same way you would debug any runs in development.