Hi GD team🙂 i'm configuring dashboard filters in Good Data cloud. I have filtered the 'Entrypoint' filter by 'Customer' filter in order to only show the entrypoints that belong to a specific customer. But I still have this 'Show all' button below and even tho it's grey it works perfectly fine and shows me all the rest (10000 to be precise which is a filter limitation). How can i get rid of 'Show all' button? It's crucial cos we're using a customer facing embedded solution and a customer can only see their own data. Thanks in advance!

Hi Masha, Thanks for mentioning this. Because indeed, Dashboard filters by themselves are not a security feature. You would need to either user Workspace data filters (WDF) or User data filters (UDF). Please note that WDFs only work per dataset. So if this cascading filter contains an attribute from a different dataset it wouldn't be automatically applied even if there's a logical connection between the dataset. This is also mentioned in our documentation:

• WDFs apply to a specific dataset and affect only that dataset.

https://www.gooddata.com/docs/cloud/workspaces/workspace-data-filters/

hey Julius, thanks for the fast reply! so we're already using a WDF for the dataset that has a customer, do you mean that for this case with entrypoints we also need to apply a customer WDF but at the dataset that contains entrypoints? or i misunderstood?

Hi Masha, my apologies, I haven't noticed that you have actually set a dependent filters in your setup. So to amend my answer - it is sufficient to have the WDF in the customer dataset if the only goal here is to restrict the default filtering for the entrypoints filter (which resides in different dataset). However, this is not a security feature and as you've correctly pointed out, the "Show all" option still appears there and reveal rest of the values.

thank you! so when you tested it did you add the exact same column to the other dataset? i read through the documentation and it seems like i need to have that exact column there to additionally map the WDF to it, is it so? my confusion comes from the fact that we're using a star schema and our entrypoints dataset is not directly linked to the customer aka doesn't have that columns there. the link to the customers is working via some intermediate tables. we have set up a WDF initially at the Customers dimension my expectation was that it would work to all the connected tables. Is it not the case of i misunderstand?

I understand what you mean. Unfortunately, this is not the case for WDFs. Only User data filters (UDFs) do work in the way that they are distributed through logical connections within the LDM. The WDFs must be set on dataset level to ensure the full security. You could either set it on some unique identifier which both datasets contain to restrict the same data rows. Or alternatively, you could create a column in your database with

wdf__

prefix (in Entrypoint for example) which would be the very same column as in the customer dataset but once you use this prefix, it only serves for the purpose of setting WDFs. It won't be available as attribute in analytics, it would be kinda "invisible" but you can map the WDF to it in the LDM and use it to filtering.

thank you, Julius, that was very helpful. Figuring out what exactly did we configured incorrectly with our WDF implementation