Forum Discussion
Query Efficiency - Stack References or Duplicate?
Hi TomDeas ,
To answer your main question: There's no difference between Referencing and Duplicating in terms of internal processing. There are some edge cases where Power Query/Dataflows will cache results, but it's safest to just assume at least one source hit per query regardless of whether duplicated or referenced.
As you rightly alluded to, the savings will be in your adjustment time. If you can create a core query that contains all generic transformations, then reference this for each subset of unique transformations, you will limit the number of areas that you need to make adjustments in.
Pete
Hi TomDeas ,
To answer your main question: There's no difference between Referencing and Duplicating in terms of internal processing. There are some edge cases where Power Query/Dataflows will cache results, but it's safest to just assume at least one source hit per query regardless of whether duplicated or referenced.
As you rightly alluded to, the savings will be in your adjustment time. If you can create a core query that contains all generic transformations, then reference this for each subset of unique transformations, you will limit the number of areas that you need to make adjustments in.
Pete
Perfect thanks Pete that's absolutely it. Was keen to take that route (stacked references) as the queries were all ultimately subsets of one another and hence helps ensure integrity of filtering later on, but wasn't sure if I was walking into a processing trap, particularly as I've had some stability issues already.
I ran overnight on basis of stacking references, and didn't encounter any issues. Perhaps my question was phrased "what's best practice here".
Thanks