Problem is in lib/matplotlib/axes/_axes.py:4969 in _pcolorargs(). It
assumes that the C argument has the .shape attribute.

        if len(args) == 1:
            C = args[0]
            numRows, numCols = C.shape
            if allmatch:
                X, Y = np.meshgrid(np.arange(numCols), np.arange(numRows))
            else:
                X, Y = np.meshgrid(np.arange(numCols + 1),
                                   np.arange(numRows + 1))
            return X, Y, C

        if len(args) == 3:
            X, Y, C = args
            numRows, numCols = C.shape
        else:
            raise TypeError(
                'Illegal arguments to %s; see help(%s)' % (funcname,
funcname))

Should be pretty straight-forward to just slap on a C = np.asanyarray(C)
before trying to access its shape.

On Wed, Oct 7, 2015 at 2:26 PM, Thomas A Caswell [email protected]
wrote:

reported via VisTrails/VisTrails#1132
VisTrails/VisTrails#1132

may be a regression from 1.3 series.

import matplotlib.pyplot as pltimport numpy as np
def func3(x, ):
return (1 - x/2 + x5 + y3) * np.exp(-x2 - y2)

make these smaller to increase the resolution

dx, dy = 0.05, 0.05

x = np.arange(-3.0, 3.0001, dx)
y = np.arange(-3.0, 3.0001, dy)
X, Y = np.meshgrid(x, y)

Z = func3(X, Y)

fails

plt.pcolor(X, Y, list(Z))# plt.pcolor(X, Y, Z) # works

—
Reply to this email directly or view it on GitHub
#5205.

We might need to do something similar for X and Y, but I know that might
have issues with units. Didn't @dopplershift come up with a cbook function
that is like asanyarray, but safer?

On Wed, Oct 7, 2015 at 2:35 PM, Benjamin Root [email protected] wrote:

Problem is in lib/matplotlib/axes/_axes.py:4969 in _pcolorargs(). It
assumes that the C argument has the .shape attribute.

        if len(args) == 1:
            C = args[0]
            numRows, numCols = C.shape
            if allmatch:
                X, Y = np.meshgrid(np.arange(numCols), np.arange(numRows))
            else:
                X, Y = np.meshgrid(np.arange(numCols + 1),
                                   np.arange(numRows + 1))
            return X, Y, C

        if len(args) == 3:
            X, Y, C = args
            numRows, numCols = C.shape
        else:
            raise TypeError(
                'Illegal arguments to %s; see help(%s)' % (funcname,
funcname))

Should be pretty straight-forward to just slap on a C = np.asanyarray(C)
before trying to access its shape.

On Wed, Oct 7, 2015 at 2:26 PM, Thomas A Caswell <[email protected]

wrote:

reported via VisTrails/VisTrails#1132
VisTrails/VisTrails#1132

may be a regression from 1.3 series.

import matplotlib.pyplot as pltimport numpy as np
def func3(x, ):
return (1 - x/2 + x5 + y3) * np.exp(-x2 - y2)

make these smaller to increase the resolution

dx, dy = 0.05, 0.05

x = np.arange(-3.0, 3.0001, dx)
y = np.arange(-3.0, 3.0001, dy)
X, Y = np.meshgrid(x, y)

Z = func3(X, Y)

fails

plt.pcolor(X, Y, list(Z))# plt.pcolor(X, Y, Z) # works

—
Reply to this email directly or view it on GitHub
#5205.

I made a stand-in for atleast_1d(), since it doesn't preserve subclasses; asanyarray() doesn't have that problem.

fails

plt.pcolor(X, Y,list(Z))

I don't see why one should expect the above to work. Pcolor and
similar functions are in displaying 2-D, so Z fundamentally has to be 2-D.

I would close this as not a bug.

plt.pcolor(X, Y, Z) # works

@efiring, it is a list of 1-D numpy arrays. Applying asanyarray() would
convert it into a 2D numpy array. I am just about done with a PR to fix
this.

On Wed, Oct 7, 2015 at 2:50 PM, Eric Firing [email protected]
wrote:

fails

plt.pcolor(X, Y,list(Z))

I don't see why one should expect the above to work. Pcolor and
similar functions are in displaying 2-D, so Z fundamentally has to be 2-D.

I would close this as not a bug.

plt.pcolor(X, Y, Z) # works

—
Reply to this email directly or view it on GitHub
#5205 (comment)
.

Fixed by #5206, I think.