SQL Pipelines

Ploomber comes with built-in support for SQL. You provide SQL scripts and Ploomber manages connections to the database and orchestrates execution for you.

graph LR ca[Clean table A] --> ta[Transform] --> m[Merge] cb[Clean table B] --> tb[Transform] --> m m --> Dump --> Plot

Feature engineering with SQL

With modern data-warehouses such as Snowflake, using SQL for feature engineering can significantly simplify the development process since the warehouse takes care of scaling your code.

You can use Ploomber and SQL to iterate on your feature engineering code quickly, then download the features to train a model using Python.

Here’s an example that applies transformations in a database, dumps a table to a local file, and then processes it using Python.

Uploading batch predictions to a database

If you’re working on a Machine Learning whose predictions must be uploaded to a database table, you can implement this with Ploomber.


Ploomber allows you to easily write SQL pipelines. Here’s an ETL example.