local data filters: option to exclude excluded items

[hogi]

☑ cool new feature
☑ could help many users!

☑ feels like a „bug“

☐ nice to have

☐ nobody needs it

Surprising, especially for new users of Jmp:

A local data filter doesn't exclude items, which are already excluded by a global data filter.
Data Filter: exclude excluded rows 

For users who work with Jmp for many years, this can be a quite logical thing.

But if you think about it objectively, it's completely unnecessary, right?

So, why?!

 In the example from the link above, a data filter which could look like this

actually looks like this (excluded, unnecessary items marked red):

A user who selects age = 13, 15,17 will not see any update in the graph; could be that he wonders what's wrong.

There is a functionality to exclude data from a graph - why not use the same functionality to exclude the items from the data filter lists?

It's already possible to apply such an exclusion manually, bu "conditional" data filters and an additional column with formula excluded().

But this requires some effort.

My proposal: a checkbox in the context menu of the local data filter: don't show data points which are excluded by the global data filter 

Then the user can decide if he wants that excluded rows are listed in the data filter (checkmark = 0) or if they should be removed (1) ...

Hm .. actually I can't think of any application case where a user intentionally decides to get excluded items listed.

But definitely, there will be some special cases ...

Can anybody give an example?

more wishes submitted by  

[danschikore]

In a Local Data Filter, the Count Excluded Rows(0) option will give the filter view that you describe.  This option is currently not supported for the Global Data Filter.  The option is also related to the Conditional(1) option, which creates a hierarchical filter view that changes the population of each filter item to exclude the rows that are selected in previous filter items.

Please let us know if these options satisfy the wish, or if we can clarify how the new wish would interact with other options that are available.

dt=Open("$SAMPLE_DATA/Big Class.jmp");
dt << Select Rows(dt << Get Rows Where(:age == 13 | :age == 15 | :age == 17 | :sex == "M"));
dt << Hide And Exclude();
dt << Clear Selection;
dt << Graph Builder(
	Size( 522, 464 ),
	Show Control Panel( 0 ),
	Variables( X( :age ), X( :sex, Position( 1 ) ) ),
	Elements( Bar( X( 1 ), X( 2 ), Legend( 3 ) ) ),
	Local Data Filter(
		Count Excluded Rows( 0 ),
		Add Filter(
			columns( :age, :sex ),
			Display( :age, N Items( 3 ), "List Display" )
		)
	)
);

[hogi]

Thanks a lot @danschikore  - wonderful, exactly what I was looking for.

I guess I always overlooked it because I never cared if excluded rows are counted or not.

Cool that this option can do so much more than just define what's counted:)

Is there an option to start a local data filter by default with 

Count Excluded Rows( 0 ) 

Would there be some disadvantage?

Why is Count Excluded Rows(1) the default setting? Speed?

[danschikore]

@hogi - I am not entirely sure of the history, but I think one possibility is just that the local filter evolved from the global filter.  The option does not make as much sense at the global level, because the filter is modifying the table row states directly.  There is some additional processing required for the option, but like Conditional(1), it may be a compelling filter setting for some use cases.

There is no preference to turn off Count Excluded Rows(0) at this time.

[hogi]

Ok, understood.

Then this wish can be reduced to 

Please add an option to the preferences to enable/disable 

Count excluded rows

by default for local data filters.