More efficient method of data load?

[twalsh0625]

Using a PBI desktop file I am pulling the Salesforce Object Data source (SF reports won't work for this due to 2000 record limit in REST API).

I am using the case object in SF, which has a TON of custom fields our admins felt were required (they weren't) and seperate work queues have seperate KPIs and methodologies for performance. I am doing multiple reports/tables with Salesforce data and have a question on efficiency of the data model.

Option 1: 

Pull the case object multiple times in query editor and apply filters to create each data table i need

Option 2:

Pull the case object once, and use the create table DAX to create the individual tables for reports and what not.

They both seem like reasonable approaches with pros/cons but which option will be less intensive, in particular to avoid data refresh (via personal gateway) timeouts?

[Greg_Deckler]

Obviously test, but my thinking is that Option 2 would be less intensive from a data load stand-point with the trade-off being more calculation in memory and DAX.