Forum Discussion

Jonathan_Oberman's avatar
Jonathan_Oberman
Icon for Quantexa Team rankQuantexa Team
2 hours ago
Solved

End to end integration tests in a project repository

Hi all - has anyone managed to successfully set up an end to end slow/integration test in their project repository?

I am hoping to achieve slow/integration tests for each of our core features (Data Fusion, QBR, GS, Assess, etc) that builds off one another. I haven't seen this in PE, just wondering if anyone has actually managed to do something similar in a project.

  • I am fully convinced CI with a controlled set of input data is the way to go. Thanks everyone

4 Replies

  • I am fully convinced CI with a controlled set of input data is the way to go. Thanks everyone

  • On our project we have setup integration/E2E tests.These run ETL, QBR and assess.I think however it is just easier to setup a CI/CD pipeline to run the stages on sample data. Its tricky to get the assess test to read all the required inputs from the ETL and QBR tests.

    This is also the recommendation from the QBR team.

  • Do you definitely want to do this as a slow test rather than just a CI job that runs all of the stages on the dev cluster?

    • Jonathan_Oberman's avatar
      Jonathan_Oberman
      Icon for Quantexa Team rankQuantexa Team

      Hey Dougall - yes, that was something I was considering too. We are using Databricks for our batch processing, in theory we could maybe use the Databricks CLI API to achieve that...