[Ingestion] Power BI Datasets and SSAS / AAS

[Daandamhuis]

Describe the solution you'd like
Since I'm working a lot with the Microsoft Stack, integrating the Power BI dataset metadata would be helpful. In terms of the metadata, this would be relatable to Looker and Tableau. There are two options, at least that's what I think.

1. Use the Power BI Admin Scanner API to read out the datasets.
2. Use PyTabular with Microsoft's dlls to read out the Tabular Object Model (TOM).

The attributes that can be retrieved are plenty, but this is a High-Level Overview;

• Measures (Expressions, Names, Descriptions, DataTypes and Translations)
• Columns (Expressions, Dataset Column Name, Source Column Name, Descriptions, DataTypes and Translations)
• Tables (Partitions, Names, Descriptions and Translations)
• Relationships within the dataset. (From Column & Table -> To Column and Table)

Downsides
Option 1 is that it can only work for Power BI Dasets (afaik)
Option 2 is that the package only runs in Windows environments (afaik). PyTabular Issue

Upside

• Option two would also make it possible to read out the metadata from SQL / Azure Server Analysis Services.
• And it would also be able to excecute DAX queries to the dataset which could be used for profiling.

Example based on Adventure Works
I've implemented PyTabular with Docusaurus for my personal website, so you can see what can be extracted.

Additional context
Measure Example

Table Example

Column Example

[Daandamhuis]

Updated the base request

[pmbrull]

Hi @Daandamhuis, sorry not an MS expert here, but is this ticket the same as being described here #9772?

[Daandamhuis]

Yes, correct 👍

[pmbrull]

Thanks, since this has some more details, I'll close the other and mark as duplicate. For ref: #9772