Analyze Data in Azure with DuckDB

Launch DuckDB

If you haven't already, download and install DuckDB on your computer. Open your terminal or command prompt, and start the DuckDB application.

./duckdb

Install and configure the Azure extension for DuckDB

Now that you have DuckDB running, you must install and activate the Azure extension. You can do this in the DuckDB CLI with the following commands.

INSTALL azure;
LOAD azure;

With the Azure extension loaded, you can configure the extension to use your Azure connection string. Use the following SET command, replacing <your_connection_string> with the value copied from the Azure portal.

SET azure_storage_connection_string = '<your_connection_string>';

You are now ready to query data files stored in your Azure container!

Query data files in Azure from DuckDB

Here is the syntax for querying a file in Azure Blob storage.

FROM 'azure://[container]/[file-name-or-file-pattern]'

For example, to query a file named survey_results.csv in a container named my_container, the SQL may look like the following.

SELECT count(*) FROM 'azure://my_container/survey_results.csv';

You can also query across multiple files with a file-matching pattern. For example, if you have separate files for each month of the year named year-month-sales.csv, you could query across the entire year using the following.

SELECT count(*) FROM 'azure://my_container/2023-*-sales.csv';

Query across multiple cloud storage providers using DuckDB

Combining the new Azure extension and the HTTPS extension, it's possible to query across multiple storage providers, should the need arise. For example, you may have historical data stored in Amazon S3 and more recent data stored in Azure and need to query across both.

-- Load and configure Azure
INSTALL azure;
LOAD azure;
SET azure_storage_connection_string = 'your-connection-string';

-- Load and configure Amazon S3
INSTALL httpfs;
LOAD httpfs;
SET s3_access_key_id='your-access-key-id';
SET s3_secret_access_key='your-secret-access-key';
SET s3_region='your-region';

SELECT t1.*
FROM (
    SELECT * FROM 's3://my-s3-bucket/sales-history.csv'
    UNION ALL
    SELECT * FROM 'azure://my-container/ytd-sales.csv'
) t1
ORDER BY "Gross Amt" DESC
LIMIT 10;

Configure your Azure connection in MotherDuck

MotherDuck provides a secure and convenient way to store your Azure connection string so you can query Azure whenever you need. To save your Azure connection string in MotherDuck, log in to your MotherDuck account and complete the following steps.

1. Click your profile menu and click Settings.
2. Under Secrets, click the ADD button.
3. Click the Secret type and click Azure.
4. Paste your connection string in the box labeled Azure storage connection string.
5. Click Save.

Query Azure data from MotherDuck

Next, create a new cell in your MotherDuck notebook. Then, write a SQL query to access your Azure storage account. For example, if you have a file saved in my_container named ytd-sales.csv, you might try the following.

SELECT * 
FROM 'azure://my_container/ytd-sales.csv'
ORDER BY "Gross Sales"
LIMIT 10;

You are ready to duck and roll with MotherDuck and Azure!