It would be hella sweet to have a way to easily track the row count after each applied step in a dataflow.
This can be useful for seeing the impact of a filter step - but most useful in seeing the impact of a join.
I've been burned many times creating what I thought were many-to-one joins only to discover that there were duplicate values in the secondary table that I didn't expect- this results in a multiplication of the records.
Example of how this would help-
I've got a table with 100,000 records. I do a left-outer join to another table where I think there is a many-to-one relationship.
I'd expect 100,000 records after this step is applied.
If there are more than 100,000 records - then I've got a many-to-many relationship that I wasn't expecting.
If I see the record count increase after this step, I can quickly identify the error of my ways and fix the problem.
Otherwise I'm left having to scan through the logic to try to troubleshoot what went wrong.
2 Comments
- miguel
Community Admin
Status changed:NewtoNeeds VotesBy design, Power Query tries to leverage lazy evaluation as much as possible. This means that we only try to compute things when necessary to not overload your data source or the compute resources need by Power Query.
While it is possible to use the "Count rows" feature from the "Transform" tab in the editor to get a count, it could be a pretty expensive operation that would take many minutes to compute depending on your data source and the resources available for the authoring session of the Dataflow.
Steve-DataCouraAdvocate I
Understood - but it would have to be faster than the "Enable details" pane - which forces this calculation plus many many more. If the process could be done on demand, rather than automatically, that would work & avoid slowing down the UI.
