Learn how to replicate data from Snowflake to PostgreSQL
Many businesses rely on online analytical processing (OLAP) solutions such as Snowflake to gain insights from data that is collected from multiple sources spread across the enterprise. Snowflake is a fully managed SaaS (software as a service) that provides a single platform for data warehousing, data lakes, data engineering, data science, data application development, and sharing of data.
As discussed in our article about data integration, moving data from across the enterprise into a centralized data analytics platform such as Snowflake provides the following benefits:
While the above benefits are significant, additional value may be extracted by moving analytical data from Snowflake back into operational systems. Furthermore, moving data back to an operational system may be beneficial if access to your Snowflake deployment is restricted.
Moving data from Snowflake into an operational system, also known as reverse ETL, is the flip side of ETL/ELT. With reverse ETL, Snowflake becomes the data source rather than the destination. Data from Snowflake is transformed to match a destination's data formatting requirements, and loaded into the destination. In other words, reverse ETL “operationalizes” analytical data by pushing it back into operational systems.
In this tutorial, you will learn how to implement reverse ETL from Snowflake into Postgres. Postgres is an online transaction processing (OLTP) relational database which is often used in operational systems. It is a robust, open-source system with a reputation for reliability and performance.
Because Airbyte supports source connectors for popular data warehouses such as Redshift, BigQuery, and Snowflake, and destination connectors for popular databases such as Postgres, MySQL, or Microsoft SQL Server, Airbyte can be used to to easily create reverse ETL pipelines.
In order to follow along the exact steps given in this tutorial, you will need to meet the prerequisites given below.
Go to Airbyte Cloud and create a new connection. Give the source a name and select Snowflake as the source type. You can read all the configuration details on the Airbyte Snowflake source documentation.
We will be creating a new database and table. Login to your snowflake dashboard and create a new database in a warehouse where you have ACCOUNTADMIN privileges. You can skip this step if you already have a database setup.
On creating the database, run the following SQL query to create a new table for our tutorial.
Our next step will be to populate the table with some sample data. INSERT the following query into our newly created table to add data.
The next steps is setting up a new source in your Airbyte Dashboard. Here is a sample configuration:
Once the configurations are complete, click Set up source.
Once Snowflake has been successfully configured as the source; you will be prompted to configure your destination. For this tutorial, we will use Heroku's managed PostgreSQL add-on to create a database host. You can skip this step if you already have a PostgreSQL host setup.
We first need to create a new Heroku app to host a new database on Heroku. Login to your Heroku dashboard and choose to create a new app
Additionally, you can change the Region to make the app's servers closer to your location.
Once you have provided the app name, click Create app.
To add a PostgreSQL database to our app, which we just created, we need to install Heroku’s Postgres Add-on. Go to Resource and look for the Heroku PostgreSQL add-on.
Perfect, we now have our database setup without any hassle. But we do need credentials to access this database.
To find the credentials and the connection URL for the PostgreSQL database, you need to navigate to the Resources tab in your app's dashboard again and select the Heroku Postgres resource
This will bring you to the configuration screen for your database. To find the db credentials, click on the Settings tab and View Credentials.
Take note of the Host, Database, User, Post, and Password; we will need these details while setting up the Airbyte destination.
To test out if we can set up a connection to our newly created database. Copy the URI provided by Heroku inside Credentials and run the following command
If everything went correctly, you should see the following output in the terminal
Now we create a new Airbyte destination using this newly obtained Postgres database, Go to Destinations in your Airbyte cloud account and click on New Destination.
Using your connection settings, you can configure your source and destination. You should be able to see the loaded schemas and tables from Snowflake.
Give the connection a name, and define a replication frequency and destination namespace. For this tutorial, we have chosen Manual as the replication frequency and the mirror source structure option.
Airbyte presents the available tables given in a warehouse. For this demo, we will only have the FLOWER table (or stream) from PUBLIC schema (or namespace).
We will set the sync mode to Incremental sync mode. In this approach, on each sync, Airbyte will only sync the new or modified data. Any changes will be recorded as a new row append to the table.
The incremental mode requires us to specify a cursor field for determining whether to sync new data or not. A typical example of a cursor field would be updated_at or modified_at columns in your database, usually timestamps. However, since our demo table FLOWER doesn’t have a field like that, we select a user-defined cursor that can function as a cursor field; one such column is FCOUNT, which contains unique incremental values for all rows.
Once satisfied with your configuration, save the connection and click Sync now to run your first sync once configured.
Once the sync is complete, you should see how many rows were copied (10 in our case). Next, we will use heroku CLI and connect to the database to view the tables created by Airbyte:
Now to find our tables that Airbyte synced, we would first have to make sure that we have access to the correct schema. Use the following Postgres command to list all available schemas in our database.
As you can see Airbyte successfully created a new table, with the name flower we can now query this table using the SELECT * operation in SQL.
As you can see, there are exactly ten rows Airbyte synced into our database.
Let’s modify one of the records in flowers and see how Incremental Appends work. Go back to your Snowflake dashboard, create a new worksheet, and run the following SQL command.
Once you update the record, Sync the table again.
As you can see, Airbyte created a new copy row with the modified values of color and fcount columns instead of updating the original record.
In this tutorial, you have learned how to:
If you have enjoyed this tutorial, you may be interested in other Airbyte tutorials or Airbyte’s blog. You can also join the conversation on our community Slack Channel, participate in discussions on Airbyte’s discourse, or sign up for our newsletter. Furthermore, if you are interested in Airbyte as a fully managed service, you can try Airbyte Cloud for free!
Learn how to easily export Postgres data to CSV, JSON, Parquet, and Avro file formats stored in AWS S3.
Learn how to stream changes from a MySQL database to Kafka using Change Data Capture (CDC).
Get your data out of Redshift and into Snowflake easily with open source data integration.