Hey. I don't understand why my workspace data filter (WDF) won't propagate downstream to my fact tables. My LDM is similar to the pic attached. I put the WDF on the organizations table - assuming that the client-group, client and all 3 fact tables will take on these values. Is GD documentation saying here that I have to apply a WDF to each of my fact tables?

The WDF is applied only to the tables that contain the specified column. If a table does not have the column, the filter is not applied, and all the data from the table is available in the workspace.

If so, then that is a huge oversight, because I will have to have the filed in each fact table - triplicating it 3 times in the LDM and thus the insight builder. Alternatively I can build 2 more WDFs at each level, but this seems overkill when it can be done with 1 WDF.

We have our wdf field in all tables that are isolated by workspace. This helps guarantee the data never gets mixed between workspace environments.

Ah I didn't see this:

You can add a prefix for WDFs to the column names in your relational model. The default prefix is wdf__. When you generate the logical data model (LDM), no labels are generated for columns named with the WDF filter prefix. If required, you can manually create and map labels to the filtered columns.

Thanks @Steve Fox

Not sure on that one. What I meant was if your WDF is like: my_account_id. By default every table that has that column in it will apply the data filter. In fact you can see that in the LDM Editor. It will show you which tables are included and excluded.

Hi @Filip Charikov, to your questions: • Each WDF is bound to one workspace and by that to one workspace hierarchy. It is possible to use WDF only in workspace hierarchy it belongs to. • After WDF creation it must be also assigned in LDM to datasets you want to be filtered. • With above being said, it is unexpected that by creating filter for dev workspace hierarchy and updating dev hierarchy model you influence executions on prod hierarchy in any way. When I check current state of dev and prod workspaces, I see that: • there are 2 WDF defined, one for prod WSs, second for dev WSs • all dev workspace datasets have assigned WDF to colulmn

org_name

• all prod workspace datasets have assigned WDF to colulmn

org_name

And my questions: • Have you succeeded to achieve setup you are aiming for? • How have you created WDF for dev WS? I cannot see any request to create it in last 7 days. What I see is update of WDF Settings for childs of dev WDF. • When you wrote

It's back to normal now, but I had to delete the WDFs from my dev workspace.

you meant dropping WDF settings related to dev WDF or dropping the whole dev WDF? From what I see, it is the first case.

@Pavel Cerny I think the issue was that my dev used to have

wdf__org_name

instead of current

org_name

and that was interacting badly with the WFDs defined

"columnName": "org_name"

.

@Filip Charikov We have identified a bug which caused your troubles. The fix should be delivered in P32. The bug manifests when you have two WS hierarchies, in each hierarchy datasets with the same datasetId and WDF to given dataset, each WDF with different column name. So exactly your setup. It can be mitigated by setting WDF column name for both hierarchies to the same value as you did.