Also include incremental refresh in dataflow for Power BI Pro users.
joao_santos4
6 years agoNew Member
A single dataflow with a single entity with few fields consumed about 3GB. In addition to taking up a lot of disk space, the Entity's load time is absurdly high. We need to optimize this so that the use of dataflow is warranted (especially if we rely on it with self-service).
Recent ideas
Publish App via REST API
The REST API allows me to programmatically deploy a new set of reports to a Power BI App Workspace, but there is no way to then Publish this App to make it available to the users via the REST API. Th...evan_wallace15 minutes agoNew MemberUnder Review4.8KViews493likes63CommentsDisable "Show data point as a table" option
The inability to disable the "show data point as a table" option is an incredibly harmful feature that poses a huge security risk for institutions trying to share out data in an easily accessible ma...eboneigh_harris2 hours agoNew MemberNew243Views15likes4Commentsactor_user_id / actor_user_name in Delta history is not captured.
The actor_user_id /actor_user_name are currently displaying a null. These are visible only when the parameter - spark.microsoft.delta.describeHistory.runtimeEnvironmentFields.enabled=true is set. But...rajorshidas013 hours agoNew MemberNew4Views0likes0CommentsFeature Proposal: Universal BI Translation and Semantic Equivalence Engine for Microsoft Power BI
Feature Proposal: Universal BI Translation and Semantic Equivalence Engine for Microsoft Power BI Category: Power BI / Microsoft Fabric / AI-Assisted BI Modernization 1. Executive Summary I propos...timster0079 hours agoNew MemberNew4Views0likes0Comments