HELP PLEASE: Filtering TWO data sets in a single box plot via Marking - only is filtering one data set!

[Byron Kelly]

Hello

I have a large data set comprised of multiple data tables that are linked via common column relationships. I am trying to plot two metrics on a single box plot, revenue and cash flow. I have created filtering relationships between all data sets and they seem to work just fine.

I would like to filter the box plot data via marking. I have set up my analysis as follows:I have a cross plot with a list of all companies and the sum of their Revenue and CF over the time period of interest. I would like to mark a company in this cross plot, and restrict the data in the box plot.

The box plot has two data sets plotted: Cash Flow and Revenue on the Y axis and years on the X. See attached images.

When I mark the data in the cross plot, only one of the two box plot metrics are filtered(CF filteres by the marking and revenue is not filtered).If i make two box plots, and plot revenue and CF seperatly, both filter when marked. I cant figure out why the combined box plot wont filter both reveneu and CFwhile the individual box plots will filter as expected.

I originally constructed the box plot by plotting CF first and then adding revenue after (so CF is the primary data set for the plot). When i do this, CF filteres when marked and revenue does not. I tried to do the opposite, plot revenue first then plot CF and when i mark the data this way, revenue will filter ok and CF will not. So I feel like im missing something.

I'll have attached a few images to explain more clearly. One is unmarked (no filtering) and the other is marked. On the marked image, only the CF data is filtered and revenue is not filtered, but I want to filter BOTH CF and Revenue by marking.

What am I missing Any help would be greatly appreciated.

Thanks

Byron

[Kirsten Smith (she/her)]

It would be helpful if you could attach your file so that we can take a closer look at the data. Make sure that the data is embedded.

[Khushboo Rabadia]

So there are two tables and you are using expression containing columns from two tables in one visualization based on column matches.

Now as per screenshot, say you have data table relations on year column, column matches would also be on year column.

Y-Axis Expression: Sum([Table1].[Revenue]) as [Rev], Sum([Table2].[CF]) as [CF] where table1 is the active/primary data table for visualization

Now you mark on company code which is another column in your table1 so table 1 gets limited by that column, Rev also updates. CF will not update because it has column matches on Year so according to that if CF in 2016 year is 250k in table 2, it will show the same. This is by design, it will show the values in plot based on the matched column

You can raise this idea in ideas portal

Below are some options:

1) you can have CF and Rev in different data tables

2) insert CF column in table 1 using insert columns based on the common columns between two tables

3) Use company code filter instead of marking to filter data in table 1 and select include filtered rows in table 2, that way box plot will be updated with filtered data from table 2

[Byron Kelly]

some more images to help explain my situation