Dataset and thin reports - how to best do dataset cleaning (naming, redundant measures/columns)

[KaidoPowerBI]

When dataset and report is combined in one Power BI report file, then it is possible to rename tables, columns and measures. The changes keep both measures referencing the columns, and visuals using the measures and columns intact, since the report knows dependencies. The problem is that this does not work, when thin reports are used, so making changes to existing columns and measures becomes a problem. One can of course always add new. I also found out, that although measure does not need table reference, then when measure is used in visual in thin report, and measure is moved to another table, then visual gets broken.

I have two wishes. One is that I could make improvements in dataset consistency - change table, column and measure names for example - which have got a bit messy after a year of developing model further and after two developers have had slightly different naming convetions. I figured one option would be to somehow combine all report pages from think report back to dataset model and later again breaking up. But I have not found easy way to do such combination. Is there a tool for this, like there exists a script to split report? Here is the script, I refer to https://powerbi.tips/2020/06/split-an-existing-power-bi-file-into-a-model-and-report/

Maybe there are some other options I could consider?

The other issue is that how in such dataset/thin reports structure  could one get overall dependency view for what measure are actually used in some visuals and what are redundant? Could something like tool from Datavizioner work here? Any experience? https://www.datavizioner.com/

[Adamboer]

It is true that when using thin reports in Power BI, renaming tables, columns and measures can be a challenge as the report does not recognize dependencies. One solution to this is to use the "Find and Replace" feature in the Query Editor to update the names of tables and columns in the dataset. However, renaming measures is still problematic as they are referenced in visuals and changing their names can cause issues.

Regarding your first wish, there is no easy way to combine all report pages from a thin report back to the dataset model. One possible solution is to export the report to Power BI Desktop, make the changes to the dataset model, and then import the report back into the thin report. This way, the updated model will replace the existing model in the report.

As for your second issue, there are third-party tools available like Datavizioner that can help you visualize the dependencies between measures and visuals. Datavizioner can generate a dependency matrix that shows which visuals use which measures, and vice versa. This can help you identify redundant measures and optimize your dataset. It is recommended that you thoroughly review the tool's features and compatibility with your version of Power BI before using it.