Hi Mike, That looks like a typo to me. Given grid arrays x(length N) and y(length M), calls to contourf(x, y, data) assume that data has a shape [M, N]. In this case, x corresponds to lons and y corresponds to lats, so the shape of the dataset should indeed be (time, lat, lon). I know it can be counter-intuitive, but it is what it is. Looking at the history, it appears that I was the one who made the typo. Sorry for the confusion!
Thanks, Alex On Mon, Mar 16, 2015 at 3:44 PM, Michael Joyce <[email protected]> wrote: > Hi folks, > > Was poking through the plotting code and stumbled upon something that I > think is just a typo but I want to get confirmation first. It seems that > the requested format for the passed dataset is (time, lon, lat) [1]. > However, the example use cases have bee using (time, lat, lon) since that > is the format that the Dataset object uses. > > I tried running an example where I took the results array from an > evaluation and reshaped the data so it was in the expected format but I > can't get it to run properly. So I assume this is just a typo, but perhaps > someone who is more familiar could elaborate? > > [1] https://github.com/apache/climate/blob/master/ocw/plotter.py#L497 > -- Joyce > -- Alex Goodman Graduate Research Assistant Department of Atmospheric Science Colorado State University
