Mysterious DAX filter terminology

[gvg]

Hi experts,

I am really getting lost in all the filter terminology. While reading different articles I come across this list of filters:

1. Context filter

2. Explicit filter

3. Implicit filter

Not to mention shadow filters described here. Leaving the latter ones for rocket scientists could you please explain on a practical example which filters are which. Say, I have a simple model:

and a PBI page:

With regard to Sales result visual:

- If have a date slicer on my PBI page - what kind of filter that is?

- What kind of filter is the Date column in the visual?

- If I filter Division by West in the Divisions table visual, what kind of filter West is in the Sales result visual context?

- If I then place filter in the Filter section of Visualizations tab - what kind of filter this is?

pbix file is here. 

Would be very grateful if someone could demistify all this.

[Phil_Seamark · Accepted solution]

By "Explicit" filters, they mean filters that are set outside the current query.

In Power BI, every visual on the report canvas is it's own query.  The query can be filtered either internally (by items on the axis of the visual, or by row/column header items), or externally (a slicer or filter that exists outside the visual).

Say you have a slicer on a report page showing countries, and you make a selection on the slicer to be USA.

Then you create a column chart visual on the same page using a calendar year on the axis and a SUM of sales as the value.

Each column in the chart probably has a different value.   The internal/implicit/query filter is applying a filter on each column for the calendar year (coming from the AXIS).  But there is also some filtering being applied by the external/explicit filter coming from the Slicer being set to USA.

If you use the ALLSELECTED function, it will remove the internal/implicit/query filtering on the calendar year, but still apply the filter coming from the slicer.  This will probably mean you will see the same value over and over for each year (which will be a total figure for USA).

Does that help?