Lessons learned from optimizing enterprise data models
summary
Shows how to mimic hybrid tables on Pro by splitting a 200M-row fact into hot (import) and cold (DirectQuery) parts and using a DAX “tipping point” (filtered aggregations) to avoid DirectQuery; cuts a chats table from 667 MB to 214 MB by rounding to hour in SQL/Power Query instead of DAX; stresses enabling manage aggregations and setting precedence so Power BI maps queries to 2k-row vs 5M-row aggregates; flips AvailableInMDX off to drop a 121 MB hierarchy to 8 MB; warns that auto date/time can create ~3M hidden dates and 176 MB overhead.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:21Data model optimization context
- 5:49Mimicking hybrid tables
- 18:20Reducing cardinality
- 27:06Implementing aggregations
- 36:34Hidden MDX properties
- 44:58Avoiding Auto Date/Time
related talks
51mOptimising DAX fusion - Power BI
51mTroubleshooting Power BI Report Performance
51mAggregations in Power BI
45mAzure DW and Power BI, reporting at the speed of light!
01h 08mAdvanced Modelling and Calculation Using the Power BI Designer
01h 04mThe Art of Prototyping Analytics
54mData Visualization with Power View and the Tabular Model
52mA deep dive into Direct Lake
