I have given up again with Dataflows. I always have problems just trying to get basic dataflows running.
My latest effort 25 CSV tables from blob storage which work fine in Power BI / Power Query. I migrated it to Dataflow and it kind of worked randomly, it doesn't most of the time, I trying to fix/change it is a nightmare. I gave up now and penned this.
If you paste your queries in from PowerBI it tries to recreate a different version of any functions that you reference. So you have to edit any ones you paste in afterward.
You can't paste all queries from Power BI because it's too big!
When you edit any entity and save it, it takes ages to "Validate" it and then crashes with a "Technical Error" that's all you get, no explanations.
Sometimes you get warnings. i.e. Table has an invalid field. Not very helpful if your table is ok in Power Query and has 90 columns.
You change the source data and edit the Power Query and it still thinks the table has an error. You have to delete the table and recreate it.
When you export the model, it's JSON but you can't manually edit it as it's rendered all the queries in one mass document.
It's all a bit on a nightmare. Every time I have tried to use dataflows (Power BI Pro) I come up against these issues and I end up abandoning the effort and just have the data model in PBI. It's a shame really as it's has got potential. I'd like to see: -
Better UI performance.
Human Readable & Editable Data Model. (programmers are quite comfortable using code).
Much better Error Reporting.
Ability to Migrate a Power Query from PBI to a Dataflow or Export as a Dataflow Data Model.
Better Paste Support, i.e. Bigger Paste Buffer, Detection & Usage of existing Functions & Table References.
Search & Replace across all objects (i.e. replacing magic numbers with parameters which is a painful experience).
Maybe if some of this stuff happens I can come back and try again and it will be fifth time lucky!
Anyone else suffering from Dataflow Blues?
Toby
