Workspace Authoring 101: Feature Caching
0 likes730 views
The document discusses feature caching, highlighting its benefits for speeding up workspace development by storing outputs of features in transformers. However, it advises against using feature caching in scenarios involving large rasters, when feature order is critical, or when utilizing variables. Additional resources and a free trial for FME are promoted for further exploration of these concepts.
Workspace Authoring 101: Feature Caching
- 2. Meet the Presenter Dave Campanas Applications Expert
- 3. Agenda What we’re covering today. ● Why Feature Caching is the coolest thing ever ● When it isn’t so cool
- 4. Type in your questions throughout the webinar. It’s also being recorded and a copy of the video, slides, and workspaces will be emailed to you.
- 5. Section 1 What is Feature Caching, and how can it help me?
- 6. What is Feature Caching ● Stores a copy of the features output for each transformer in workspace ● Workbench setting - is not saved in workspace ● Does not apply when workspace run in Quick Translator, FME Server, or command line FME
- 7. Demo
- 8. Feature Caching can speed workspace development By reducing the number of times the full workspace must be run, and automating the exposure of new attributes and values.
- 10. Section 2 When should feature Caching not be used?
- 11. Raster workflows ● Rasters are usually quite large ● A copy must cached to disk for each transformer ● FME optimizes raster handling for best performance ● Caching disrupts raster optimization When working with rasters: ● Disable Feature Caching ● Bookmark and collapse raster section in mixed raster/vector workspaces
- 12. Workflows using Variables ● VariableSetter and VariableRetriever often used to efficiently transfer information between feature streams and to create lookup tables ● Variables stored in memory and not cached with features ● Variable state lost in partial workspace runs
- 13. Workflows where feature order is important ● FeatureMerger, Clipper offer performance options if feature order can be controlled ○ ie. Clippers First option ● Order may not be the same from caches of multiple streams
- 14. Resources FME Evangelist Blog, with more examples: https://www.safe.com/blog/2018/05/caching-data-fme-evangelist174/ Knowledge Article: https://community.safe.com/s/article/feature-caching-and-performance
- 15. Summary Feature Caching can help you develop workspaces faster. Feature Caching should not presently be used when feature order is important, or when variables are used.
- 16. Try it Yourself! Download FME 2021.0 Download the workspace from this webinar Apply it to your own data to manage changes Free Trial | Upgrade
- 17. Thank you! Get a free trial of FME at safe.com/trial Type in your questions for Q&A