Hands-On Tutorial: In-Database Charts


This lesson is part of a series of “Usage of SQL and Dataiku tutorials” that begins with the Hands-On Tutorial: Sync Recipe.


This content is also included in a free Dataiku Academy course Dataiku & SQL, which is an optional part of the Core Designer learning path. Register for the course there if you prefer to track and validate your progress alongside concept videos, text summaries, hands-on tutorials, and quizzes.

When we create charts on database datasets, the summary statistics used to build the charts can be computed in-database.

  • In the customers_enriched dataset, click the Charts tab.

  • Drag pages_visited_avg to the X axis and Count of records to the Y axis in the default bar chart.

At the moment, the chart is based on a sample of the data.

  • Navigate to the Sampling & Engine tab of the left pane.

  • Select In-database as the execution engine and click Save.

In this case, the shape of the bar chart is essentially unchanged, but when the sample is not representative of the full data, using the full dataset can be illuminating, and using the in-database engine can speed your results.