Allow a TERR data function to output columns straight into a visualisation

Currently, TERR can output columns only into a fully loaded data table. If we wish our TERR column result to respond to filters or markings*, we must import not just the filtered table but also at least one column of the potentially massive original data table and pass it into TERR, merge it with the small results we are interested in and pass that long column right back and add that whole column to the Spotfire table to then depict on the visualisation only the small filtered results. This makes the visualisation refresh much slower than it would.

Of course we could have used a TERR expression, but the writing of these are so much more constricted than a full script that some calculations are virtually not possible, such as I believe the one in the e.g. below..

* e.g. to calculate the ttest of the difference of means between two points in a scatterplot that depict a mean,