Loading an existing dataset/semantic model to Lakehouse

[PowerNewUser]

I have a dashboard that has been created by another department and I need to pull some of the data from it into our lake house. We are currently doing it manually where we export the data of a visual and import it in using a dataflow gen2. 

I have been given contributor access to the other department workspace - however have not found a way to pull that data into my lake house. Any suggestions? I tried the dataflow connector but dont see the dataset in question in my list (it is a regular power bi dashboard importing data from a source and published to a premium workspace)

[GilbertQ]

Hi @PowerNewUser 

You can use a notebook to query data from another parvia semantic model and then load that into your lake house Here is an example. Fabric Semantic Link and Use Cases

[PowerNewUser]

@GilbertQ  Seems to be promising - however am not experienced with the notebooks/PySpark - will try. Any other direct alternatives?

[PowerNewUser · Accepted solution]

Adding what I did for reference- 

1. added a notebook with following code 

# import fabric
from sempy import fabric as FabricDataFrame
# read the table 
df_tables_ReqInfo = FabricDataFrame.read_table(workspace ="....." , dataset="....." , table="...")
# drop any extra columns
df_tables_OffersHireData.drop(['....', '.....'], axis = 1, inplace=True)
# write to lakehouse
df_tables_OffersHireData.to_lakehouse_table(name="....", mode="overwrite" )

2. Created an enviroment to preload the semantic-link library and set it to as default for workspace - this is required if you want to call the notebook in a pipeline as pip is not allowed at that time. 

[mkmorrison]

Doew the dataframe name change from "df_tables_ReqInfo" to "df_tables_OffersHireData" have any effect?

[PowerNewUser]

thats just a typo..I had two tables I was processing and ended up removing the middle steps..was trying to explain the solution without exposing my table/column details. 

[Luci7]

@PowerNewUser Thank you so much for posting the solution. I was trying to find a way and you did it! I'll add a link to your solution in the post I created.