Converting a calculated table into an import mode table

[reubwork]

For slicers, I often use a pattern whereby I create a calculated table that does a distinct across the slicer columns from some import table. 

```

Slicer Table = 
DISTINCT(
  SUMMARIZE(
     FactTable,
     FactTable[SomeColumn]
  )
)

However In one of my larger datasets is taking a bit too long to refresh, so I'm looking to simply the model by replacing the above with an import query from the datasource. 

I can do this the manual way, - delete the old table, create the new, recreate relationships, update any measures and then fix any visuals.... but was wondering whether anyone knows of a technique that doesn't then require fixing all the visuals. If I could switch the import partition to an M query... everything downstream would remain the same as it has the same column names, lineage... 

I've tried updating the partition in the tmdl definition, but power bi complains that you can't convert a calculated table to an import table. 

[reubwork]

nb: the difficult bit isn't creating the table and remapping the relationships. It's that it breaks any visuals using the original table columns. Ideally there'd be a way to swap in a new version of the table object, with the same columns, and types... and have the downstream report visuals not notice anything had changed. Logically, I think this makes sense but the implementation seems to not allow. 

[Idrissshatila · Accepted solution]

Hello @reubwork ,

I tried creating the dim table i POwer query but it didn't work to replace the calculated table, so you need to delete it and then re-create the relationships again.

sorry bro I tried!

[reubwork]

Yeah, looks like it. I'm wondering why though, under the hood.