Pivot++

[jib]

I am thoroughly enjoying the pivot command. A few aspects that could make it even more powerful:

a) I often use the “Select underlying rows” from a r-click on the bar and copy the data to a spreadsheet to do some temporary check/computation. Instead of having to r-click a bar to get to this function, could this be made to work from an r-click on the applicable row in the little data table also? (Often my pivot/bars are hidden).

b) It would be very nice if we could extract more than a single calculation from a pivot. Often I like a table with min/max/mu/sigma, one of which I may plot with the pivot’s bars though most of the time I just extract the data to rows/columns to use a line plot or other computations before plotting. Right now this takes multiple pivot commands that are essentially the same and only switch the Calculation parameter.

c) I end up having quite a list of pivots in some of the projects. It would be nice to be able to give them a description that can be read when the command is collapsed. Right now I have to open each to see and digest the set up for a few seconds to figure out which one I want to work on.

[richard]

I would like to add that I also enjoy the pivot command. I typically use it to sort data and apply masks/filters to it. Then I export the flattened data and use it to make plots. Most of the time these are then scalar plots.

When I want to add new data to the dgraph file, I add a new pivot command etc. However, it would be very handy if I could duplicate a pivot such that its export (Flat Value, Flat Column, Flat Row data) is duplicated with it and as such I can plot the (pivot-sorted) new data immediately. At the moment I have a new pivot, I need to export the flattened columns again, make a new scalar plot, select the flattened columns; which is a pain when a lot of data is to be read in.