Allow the Row State column to have multiple Row State settings

[txnelson]

I would like to see the Row State column to have multiple row state settings available at the same time.  Such that Platforms that are run against the data table would be allowed to select which row state copy it should use.  

Use Cases

1. Open the Big Class data table, run twice Bivariate using height by weight.  Color the data points in one chart based upon row state colors based on Sex, and the colors in the second Bivariate based upon the second row state settings where the colors are based upon age.
2. Open multiple platforms.  Set different platforms to different row state versions.  Change the selections in a particular platform changes the display in only the platforms that have been set to use the same row state version.
3. Open the Subset platform, and select which row state version to use for the selection of the rows to be copied to the new data table.  The Link To Original needs to honor which row state version is associated to the selection when the subset was created.

[gzmorgan0]

@txnelson This is a cool concept, but whew, what a large change:

• Each platform would require a new role, row state. ( Y, X, By, Freq, Weight, Row State)
• I assume you would want the specified row state column to be dynamically linked to the associated platforms, so if multiple bivariates, or variability charts are assocaited with a specific row state column, a selection or exclusion in one of those platforms would update that row state, as well as, all platforms associated with that row state. [?]

An alternative might be an interactive (UI)  menu item to create a new context, where the context allows the user to add graphs and analyses.

[txnelson]

I don't think it would be a "Large" change.  Yes the dialogs for each of the platforms would have to be changed, but the the platforms themselves would remain the same.  It would be the routines that access the row states that would have to be modified, so that when a platform is run, the data table routines would just feed back to the platform, the information.

Today, JMP can do my suggested enhancement, but it requires multiple copies of the data table.   With data tables in the millions of rows, one does not want to do that.  There are other software platforms that have similar functionality to what I am suggesting, and they do it because of the large size of the data, and the desire not to have multiple copies of the data floating around.