Power BI Data Modeling Question

I have 2 questions on my data modeling:

For my Power BI report requirement I had to integrate 2 data sources of different grain

Student Data Source(176 columns) - Student identifier, term, program,major, major school, major department, student level, type ,demograhics - ethnicity,citizenship,country origin, sex - Min 1 row/per student to a max of 3 rows based on the major count and program count for each term

Course Data Source(315 columns) - Student identifier, term, course attributes - course number, course section, instructor, active course details,credit hours - Min of 4 row/student to 6 rows for each term

Additionaly, both the datasources have snapshots captured everyday. So for example for each snapshot, the student data source has 9000 records and course data source has 45000 records roughly. Both the tables has 3 years of data and roughly the data count is around 35 million.

Due to the nature of the grain if I combine these flattened tables, it becomes a cartesian join and the performance would be worse. So following are the steps taken:

1. Removed the unwanted columns from the silver layer and moved the refined tables to gold layer:

Student Data Source(81 columns)

Course Data Source (119 Columns)

2. Created bridge tables as below to filter both the data sources on a common slicer

Dim School - union of distinct schools from both the data source

Dim Department - Union of distinct department from both the data sources

Dim Student Level - Union of distinct Student Levels from both the data sources

Dim Term - union of distinct terms from both the data sources

I used the onelake mode semantic model instead of import or Direct query, due to the number of columns and the multiple snapshots.

Earlier, when I tried with import mode, I had delay issues in rendering the data to the report, whenever I made any changes to the datatype/renaming the values,

etc.

It took 45 minutes per change. So I opted for the Directlake mode, which helped in quick data rendering.

Question 1 - Is my approach correct?

Question 2 -

2.1 The dims I created are having blanks, but if I ignore the blanks, my totals match.

2.2 Although it is a union, I see some values missing.

Example: Student Data Source has an additional department Healthcare Management offering a Major, but that department is not in Course Data Source, as they don't offer any courses

when I list the department in a separate slicer, I see the Healthcare Management, but if I add them in a matrix adding school(from dim_school) and the department(from dim_department), and select the Student enrollment metric(built using student data source), healthcare management is missing. What could be the reason?