accessing public filters in cubes

I have a cube which contains a python transform. So it has [input]->[python]->[output].

I need to allow users on the dashboard to filter the cube before the python transform. The python does things to the data, but the columns that I need for the filtering are removed at this stage due to aggregation and similar calculations.

What I have done is do make output parameters on the input transform, and make these public. Is this the right way to do it?
I can’t work out how to access these from the dashboard, as the necessary columns are not in the output transform.


would this work?

I realize this assumes the input is from a database.

1 Like

All element output parameters, if made public, end up as metric set inherited parameters. You can see them in the UI when defining view parameters.

Note that bridges parameters cannot handle tokens (eg. last month). This is a regression from version 6 to 6.1 if I recall well, and a massive drawback.

No exactly a regression, but a feature requested by customers. Multiple customers asked us NOT to process token automatically, but let the user handle them in scripts.

I have been using placeholders and setting them to public. Then on the dashboard, depending on the type of placeholder and the type of filter used, you can connect the filters and placeholders together.

Some scripting is used at times to, for example, have a select’s selected option fill in in a string type placeholder.