Earn a 50% discount on the DP-600 certification exam by completing the Fabric 30 Days to Learn It challenge.
I often find I have a need to create composite keys to get proper relational joins in the data models.
I see this a lot with GUIDs and datetimes needing to be joined together when reviewing SPO data.
Is there a "ideal" way to generate a unique key to join on, does PowerBI's data model care what the keys it joins on look like?
I guess it would require generating the new surrogate key during data load with the intended benefit that having a single column of INT data type would be more efficient than a combined column that contacinates a GUID and a datetime stamp, perhaps that is not actually that big a deal for performance?
Hi @OneWithQuestion,
>>I guess it would require generating the new surrogate key during data load with the intended benefit that having a single column of INT data type would be more efficient than a combined column that contacinates a GUID and a datetime stamp, perhaps that is not actually that big a deal for performance?
I think these operation may not has big effect to improve performance.
On my opinion, you can split tables to different part and link wiht bridge tables to reduce redundancy and complexity, it should more effective than what you mentioned.
BTW, tables who has complex structure and huge amount of columns will effect processing performance.(include calculation, loading and visualization performance)
Regards,
Xiaoxin Sheng
User | Count |
---|---|
101 | |
90 | |
80 | |
71 | |
69 |
User | Count |
---|---|
114 | |
98 | |
97 | |
73 | |
72 |