1 2 Previous Next 16 Replies Latest reply on Oct 11, 2014 4:29 PM by Matt Lutton

# Dynamic Exclude Problem

I have a scatter plot where a parameter allows the user to select a field to use for the color of the bubbles on the scatter plot. The problem is that when a user selects one field to color by and excludes data, and then changes to a different field to color by, the excluded data comes back.

In the attached example, select a few points and exclude them, then change the bubble color from country to city. You'll notice your excluded data has returned. I want to prevent that from happening. Any ideas how this can be achieved ?

• ###### 1. Re: Dynamic Exclude Problem

No it doesn't. You're excluding countries, right? Then you're switching to city, right. And you're thinking the your excluded marks have returned. But now switch back to country. Those countries are still missing. What you have is completely separate sets of marks depending on which parameter you've chosen. And while Tableau is smart, it really can't know that certain cities are in certain countries, unless you tell T.

--Shawn

• ###### 2. Re: Dynamic Exclude Problem

No I am not excluding countries, I am excluding individuals. Each circle on the scatterplot represents one person. If I exclude on person from Canada(not all Canadians) and then switch the bubble color to City, that person will come back, which to me makes no sense. It's as though Tableau is creating 3 rows out of 1 for that person, one for each possible value of bubble color. The thing is I want to exclude the person and bubble color isn't necessary to identify the row to exclude.

• ###### 3. Re: Dynamic Exclude Problem

Thomas select Country. Now look in the lower left corner and you'll see this:

So there are 39 marks in the view. You've colored them by Bubble Color calc and labeled them with names. Without excluding any marks change Bubble Color to City. You still have 39 marks. Same for Dept. Right?

It's not that Tableau is drawing '3 rows out of 1 for that person', it is that when you put the Name field into the view you are telling Tableau to slice/dice all of your marks by Name, which gives you 39 marks. To see this set your parameter to Country and then drag the Names field out of the view. How many marks do you have now? Two, right?

Hit back button. Now set your Bubble Color parameter to City, and again drag the Name field off of the Text shelf. Now you have 7 marks.

Essentially this is all about granularity, or level of detail, and understanding what that means, and what (and how) that tells Tableau to behave. While you are deleting marks in one scenario (setup), you are not deleting those marks in every possible scenario; at least not the way you have your data/viz set up.

FYI: Using the Exclude/Keep Only options are less than ideal, and can lead to blind alleys your users might not know how to back out of.

Cheers,

--Shawn

1 of 1 people found this helpful
• ###### 4. Re: Dynamic Exclude Problem

There are 39 rows in my data set, by using Name in my Viz I am operating at the most granular level of detail. Yet, because Bubble Color is dynamic, Tableau is essentially acting like I have 117 rows of data.

In fact, if I start with Country as my bubble color and exclude all my obs, then switch to City and do the same, and then switch to Dept and do the same, when I look at the Exclusions Filter that has been created it has 117 values in it.

My question really is how do I prevent that ? Can I prevent it ? Is there a way to tell Tableau that in effect Name is a primary key, and if I exclude a primary key value, exlcude it everywhere ? Or is there a way to tell Tableau to ignore the bubble color field when it creates the Exclusions filter ?

• ###### 5. Re: Dynamic Exclude Problem

Sorry, I obviously misunderstood your need. Hopefully others will know how to solve your issue.

--Shawn

• ###### 6. Re: Dynamic Exclude Problem

Thomas, just curious: Did you start out as a QlikView user?

--Shawn

• ###### 7. Re: Dynamic Exclude Problem

No, I've never used QlikView

• ###### 8. Re: Dynamic Exclude Problem

No problem, thanks for taking some of your time to look at it and offer suggestions, I appreciate it.

• ###### 9. Re: Dynamic Exclude Problem

Thomas thanks very much for the reply. I only asked because I've been coming across the sort of behavior you seem to be expecting of Tableau from folks who have used QlikView. QV has a filter-hierarchy feature similar to what you seem to have been expecting. It is an interesting concept -- but I'm not sure it would/will 'fit' into the Tableau way of thinking.

Cheers,

--Shawn

• ###### 10. Re: Dynamic Exclude Problem

Turn the Bubble Color Display dimension into an Attribute or any other aggregation (MIN, MAX, AVG). This won't have any effect on the visualization because the granularity of the view is dependent on the Name dimension, and then the built-in click+drag to Exclude functionality will only be using the Name dimension.

There's also an additional workaround using a self-Filter Action where the Filter Action would filter that worksheet.

Jonathan

• ###### 11. Re: Dynamic Exclude Problem

So simple, yet I'm not sure I would've seen this solution immediately.  Thanks for sharing Jonathan.

• ###### 12. Re: Dynamic Exclude Problem

Thank you so much, this is great.

• ###### 13. Re: Dynamic Exclude Problem

Hi Matthew,

Honestly, I didn't see it right away either. I worked on the Filter Action idea first because I'd just watched a TC14 video that talked about self-filter Actions (I think it was Marc Reuter's Jedi talk, maybe Andy & Bryan's FB presentation). Anyways, the underlying key (that Shawn was alluding to) is thinking about the granularity of the view. In a table calc context I'm always thinking about what granularity is absolutely required because anything more than the minimum can potentially create further complications. In this context where it's just a couple of dimensions I let myself not do that first when I really should have.

Jonathan

• ###### 14. Re: Dynamic Exclude Problem

Good to know; this is a really good example of understanding "the granularity of the view", and how that impacts what Tableau is doing when we interact.  And a good example of how easy it can be to overlook a simple solution!

1 2 Previous Next