Power BI data loss in workspace while using GIT

[August_Stidberg]

Hey, we are working with Power BI and GIT, which we think works well for version control, etc. We use both deployment pipelines and GIT. We have two workspaces: TEST and PROD. When we finish testing, we deploy it to PROD.

(In the long run, we plan to use only GIT, but we want to ensure it works flawlessly for us before taking any risks.)

One problem I’ve been encountering is that when i change stuff in the query for a table and commit it through the GIT process. This forces us to do a full refresh on the dataset, which is both time- and data-consuming. Our feeling is that this happens a lot more now with the GIT process compared to the deployment pipeline setup (DEV → TEST → PROD).

Does anyone else have this problem? Is there a solution for it?

As things stand, we’re considering going back to the pipeline process, but if we could solve this issue, we see many more benefits with GIT

Thanks in advance 

[lbendlin]

Why would this be a data loss?

Like in deployment pipelines you can use (git excluded) parameters to control how much data to fetch in each stage/workspace.

[August_Stidberg]

Perhaps "data loss" isn't the most accurate term for this issue. The main problem is that previously, when using the basic pipeline approach, we could make changes to a query for a single table in a semantic model and then refresh only that specific table using Tabular. A full refresh wasn’t necessary.

However, with the current Git-based approach, if we make a change to a query for any table, all the data in the dataset gets wiped, forcing us to perform a full refresh to restore it.