Member-only story

Tableau CRM Performance Benchmark: Data Prep Recipes vs Dataflows

Pedro Gagliardi
4 min readFeb 28, 2022

Last week I started working on updating my previous guides that used Dataflows to build the datasets so they can use the latest features from Recipes and Data Prep.

While I was doing so, I remembered that on a Learning Days Webinar it was mentioned Data Prep Recipes are faster than Dataflows because they use a new and improved Spark infrastructure in the backend. So I went ahead and compared the run time of two of my Dataflows vs Recipes.

The first test was on the “Child to Parent” Dataflow that shows how you can roll up any field at a lower grain to a higher grain. I recreated the Compute Relative calculations and the rest of the nodes using the exact same logic so the comparison would be fair.

This is the Dataflow version:

Original Dataflow nodes

This is the Recipe:

The author made this story available to Medium members only.
If you’re new to Medium, create a new account to read this story on us.

Or, continue in mobile web

Already have an account? Sign in

Pedro Gagliardi
Pedro Gagliardi

Written by Pedro Gagliardi

Einstein Analytics geek since 2015… I know, it was called Wave back then!

Responses (2)

Write a response

Pedro, I just finished a report for my client comparing processing times from Dataflows to Recipes. Since they have a +100M rows dataset, the time differences are bigger, a Dataflows that last minutes, using recipes takes hours.
I also tested the…

I am super concerned with this as dataflows aren’t highly performant to begin with compared to other workflow tools.

The node breakdown does help and register being the biggest item as we all know it stores data in a columnar indexed format.

Interested to see how this improves