In working on code for matplotlib, I would generally find it easier to provide an array of objects than to provide an array of X values and a second array of Y values. I understand that X[] and Y[] is the way that matlab does it, but we would find it easier to provide an array of either tuples [(x1,y1), (x2,y2), ...] or else an array of objects with .x and .y properties.
Other than doing our own cover, is there any other way to do this? Any thoughts of adding this functionality?
-Simson
Simson Garfinkel wrote:
In working on code for matplotlib, I would generally find it easier to provide an array of objects than to provide an array of X values and a second array of Y values. I understand that X and Y is the way that matlab does it, but we would find it easier to provide an array of either tuples [(x1,y1), (x2,y2), ...] or else an array of objects with .x and .y properties.
Other than doing our own cover, is there any other way to do this? Any thoughts of adding this functionality?
Internally, we often do end up using Nx2 ndarrays.
I don't think it makes sense to complicate the API by directly supporting other function signatures, however, if this is what you are proposing. For example, we have plot(y) and plot(x,y); to support either of the alternatives you mention above would require either a kwarg or a different function name. Personally, I would rather leave it to the user to take care of this as needed, and keep the additional complexity out of matplotlib.
Eric
-Simson