I am striving to improve my workbook performance as much as possible. One way I am doing so is reducing my overall extract size. I recently did a comparison:
112 MB - Extract
57.9 MB - Extract (with Extract Filter on one dimension)
25.1 MB - Extract (no Extract filter, but dashboards using only 3 dimension, all others hidden)
13.7 MB - Extract (with Extract filter on one dimension, dashboard using only 3 dimensions, all others hidden)
As you can see, I was able to reduce my sample extract by ~ 88% !
However, I have two follow up questions I am hoping someone on here can definitively answer:
1. If I publish an extract as a data source, does it auto hide unused fields and thus reduce the extract size?
2. What does the "Optimize" feature do?
1. No, publishing does not auto hide anything in your datasource. It goes as is.
2. Optimize embeds certain calculation results into extract so they can just be used and not recomputed on the fly. Plus it might remove materialized calculations from extract if those are no longer needed or valid.