Hi @Gopal_PV
Thanks for reaching out to the Fabric community.
To get all of your Snowflake views working together smoothly in Power BI, you can reshape them into a classic star schema:
Create a proper Date dimension:
- In Power Query, point at your Calendar Date view.
- Rename the date column to DateKey (type “Date”) and keep only the extra fields you actually need (Year, Month, MonthName, DayOfWeek, etc.).
- Load that table and, on the Modeling ribbon, mark it as your Date table.
Build one Account lookup table:
- Pull in each of your metric views (Units, Screening, Subscription, Payments, AAS, TotChargesRevenue).
- In each query, remove all columns except ACCOUNTID and NAME.
- Append them all together, remove duplicates, and call the result DimAccount.
Turn each view into its own fact table:
- In the Units query (and repeat for the others), add a custom column:
DateKey = #date([YEAR], [MONTH], 1)
- Hide or delete the original YEAR/MONTH columns if you like, and rename your measure columns (e.g. NewUnits, ExpansionUnits).
- Load each one into the model under a new name (FactUnits, FactScreening, etc.).
Hook up your relationships:
- In the diagram view, drag DimDate[DateKey] onto each fact table’s DateKey (one-to-many, single-direction).
- Then link DimAccount[ACCOUNTID] to each fact table’s ACCOUNTID.
- Every relationship should flow from dimension → fact, not the other way around.
Build your visuals off that star:
- Always use fields from your dimensions (Year, Month, Account Name) on your axes, slicers, or filters.
- Pull your measures (for example SUM(FactUnits[NewUnits])) straight from the fact tables into charts or cards.
With this pattern you’ll avoid accidental inner joins, keep every row of your metrics, and be able to slice by date or by account without lookup failures. Let me know if you’d like any sample M-code or DAX formulas to get it working!
Thanks
Cheri Srikanth
CST_Community Team
