Something like reshape but accounting for missing values

I have a columnwise array of x- and y-coordinates with values at those x,y pairs. The full set of coordinates is predetermined and known, but in this array there isn't a value at every x,y pair. When there's no value, it's just skipped, so the original array cannot be rearranged using reshape. For example, the original array may look like this

x1 y3 v13

x2 y2 v22

x2 y3 v23

x3 y1 v31

x3 y2 v32

x3 y3 v33

and I want this

NaN NaN v31

NaN v22 v32

v13 v23 v33

I want to rearrange this original array to put the values in an array whose size is determined by the full set of coordinates with NaNs where there are missing values, and I want it to be fast. It'd be kind of like reshape, except, because there are missing (not empty) values, I don't want to cut up the column in a regular way to reshape. I think if there was a 2D interpolation option that allowed one to replace the "missing" elements with a constant, that would work, but I can't find that. I would replace them with NaNs in that case.

Answer 1

[Image Analyst]

Perhaps fillmissing()?

Or scatteredInterpolant()? (Dee attached demo.)

[Daniel]

fillmissing needs an element there to replace essentially. I have missing elements, so I have to first have a way to say where to fill. scatteredInterpolant would work if you could "interpolate" by putting a constant value wherever interpolation was required. In my case, I would put NaNs there.