TECH6A: Testing data – golden vs. generic expected

September 28, 2022 from 15:15 to 15:35

Speaker: Michal Pilarski, Jeppesen (PL)

During his career, Michal has been always connected with geospatial data and GIS geoprocessing. He likes to find and overcome challenges in Testing Big Data with geometry attributes. He has experience in preparing the testing strategies for ETL systems that extract, transform and load massive geospatial data. His technology stack is related to Python, Pytest, ArcGIS, QGIS, FME, Robot Framework, HP ALM, and Geopandas.

The main goal of Michal’s lightning talk is to compare two approaches to expected results preparation for testing data. The first method Generic Expected is fully automated and defined as generating expected results based on an independent and alternative testers system that is fully validated by humans and unit tests are applied. The second method Golden Expected is semi-automated and defined as generating expected results based on the developer’s system where expected results are fully validated (corrected if applicable) by humans via matching proper input test data with output expected data. Which approach is better? Let’s find out together.