Thread: [Matplotlib-users] Setting y-tick locations on an image plot changes y-axis limits, Bug or expected

Hi,

I've recently noticed that setting the y-tick locations on an image plot
changes the y-axis limits, while changing the x-tick locations does not
change the x-axis limits.  I wouldn't have expected either to change the
axis limits, but it seems quite inconsistent that the y-axis and x-axis
behave differently in this situation.

The axis limits are easily restored by a call to "ax.axis('image')", but
this seems unnecessary.

A very quick search of the bug tracker doesn't reveal anything matching this
description...

Is this expected behavior or a bug?

As an example, setting the y-tick locations changes the y-axis limits:
import numpy as np
import matplotlib.pyplot as plt

num = 10
data = np.arange(num**2).reshape((num,num))

fig = plt.figure()
ax = fig.add_subplot(111)

cax = ax.imshow(data)

ax.set_yticks(range(num//2))

fig.show()

However, setting the x-tick locations does not change the axis limits...
import numpy as np
import matplotlib.pyplot as plt

num = 10
data = np.arange(num**2).reshape((num,num))

fig = plt.figure()
ax = fig.add_subplot(111)

cax = ax.imshow(data)

ax.set_xticks(range(num//2))

fig.show()

As a reference, I'm running matplotlib 1.0.0 with python 2.7.0 on a 64-bit
Linux (kernel 2.6.31) machine.  I can confirm this on python 2.6 and
matplotlib 0.99 as well, though.

If this isn't expected behavior, I'll be glad to open a bug report.

Thanks!
-Joe

On 08/20/2010 12:18 PM, Joe Kington wrote:
> Hi,
>
> I've recently noticed that setting the y-tick locations on an image plot
> changes the y-axis limits, while changing the x-tick locations does not
> change the x-axis limits.  I wouldn't have expected either to change the
> axis limits, but it seems quite inconsistent that the y-axis and x-axis
> behave differently in this situation.
>
> The axis limits are easily restored by a call to "ax.axis('image')", but
> this seems unnecessary.
>
> A very quick search of the bug tracker doesn't reveal anything matching
> this description...
>
> Is this expected behavior or a bug?
>
> As an example, setting the y-tick locations changes the y-axis limits:
> import numpy as np
> import matplotlib.pyplot as plt
>
> num = 10
> data = np.arange(num**2).reshape((num,num))
>
> fig = plt.figure()
> ax = fig.add_subplot(111)
>
> cax = ax.imshow(data)
>
> ax.set_yticks(range(num//2))
>
> fig.show()
>
> However, setting the x-tick locations does not change the axis limits...
> import numpy as np
> import matplotlib.pyplot as plt
>
> num = 10
> data = np.arange(num**2).reshape((num,num))
>
> fig = plt.figure()
> ax = fig.add_subplot(111)
>
> cax = ax.imshow(data)
>
> ax.set_xticks(range(num//2))
>
> fig.show()
>
> As a reference, I'm running matplotlib 1.0.0 with python 2.7.0 on a
> 64-bit Linux (kernel 2.6.31) machine.  I can confirm this on python 2.6
> and matplotlib 0.99 as well, though.
>
> If this isn't expected behavior, I'll be glad to open a bug report.

It's a bug.  I think I have it fixed, but I need to do some more 
testing, so I won't commit anything tonight.

Eric


>
> Thanks!
> -Joe
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On 08/20/2010 12:18 PM, Joe Kington wrote:
> Hi,
>
> I've recently noticed that setting the y-tick locations on an image plot
> changes the y-axis limits, while changing the x-tick locations does not
> change the x-axis limits.  I wouldn't have expected either to change the
> axis limits, but it seems quite inconsistent that the y-axis and x-axis
> behave differently in this situation.
>
> The axis limits are easily restored by a call to "ax.axis('image')", but
> this seems unnecessary.
>
> A very quick search of the bug tracker doesn't reveal anything matching
> this description...
>
> Is this expected behavior or a bug?
>
> As an example, setting the y-tick locations changes the y-axis limits:
> import numpy as np
> import matplotlib.pyplot as plt
>
> num = 10
> data = np.arange(num**2).reshape((num,num))
>
> fig = plt.figure()
> ax = fig.add_subplot(111)
>
> cax = ax.imshow(data)
>
> ax.set_yticks(range(num//2))
>
> fig.show()
>
> However, setting the x-tick locations does not change the axis limits...
> import numpy as np
> import matplotlib.pyplot as plt
>
> num = 10
> data = np.arange(num**2).reshape((num,num))
>
> fig = plt.figure()
> ax = fig.add_subplot(111)
>
> cax = ax.imshow(data)
>
> ax.set_xticks(range(num//2))
>
> fig.show()
>
> As a reference, I'm running matplotlib 1.0.0 with python 2.7.0 on a
> 64-bit Linux (kernel 2.6.31) machine.  I can confirm this on python 2.6
> and matplotlib 0.99 as well, though.
>
> If this isn't expected behavior, I'll be glad to open a bug report.

Joe,

Thanks for the report.  The fix is in svn 8652 and 8653.

Eric

>
> Thanks!
> -Joe

Thanks for the amazingly quick turnaround!
-Joe

On Sat, Aug 21, 2010 at 1:34 PM, Eric Firing <ef...@ha...> wrote:

> On 08/20/2010 12:18 PM, Joe Kington wrote:
> > Hi,
> >
> > I've recently noticed that setting the y-tick locations on an image plot
> > changes the y-axis limits, while changing the x-tick locations does not
> > change the x-axis limits.  I wouldn't have expected either to change the
> > axis limits, but it seems quite inconsistent that the y-axis and x-axis
> > behave differently in this situation.
> >
> > The axis limits are easily restored by a call to "ax.axis('image')", but
> > this seems unnecessary.
> >
> > A very quick search of the bug tracker doesn't reveal anything matching
> > this description...
> >
> > Is this expected behavior or a bug?
> >
> > As an example, setting the y-tick locations changes the y-axis limits:
> > import numpy as np
> > import matplotlib.pyplot as plt
> >
> > num = 10
> > data = np.arange(num**2).reshape((num,num))
> >
> > fig = plt.figure()
> > ax = fig.add_subplot(111)
> >
> > cax = ax.imshow(data)
> >
> > ax.set_yticks(range(num//2))
> >
> > fig.show()
> >
> > However, setting the x-tick locations does not change the axis limits...
> > import numpy as np
> > import matplotlib.pyplot as plt
> >
> > num = 10
> > data = np.arange(num**2).reshape((num,num))
> >
> > fig = plt.figure()
> > ax = fig.add_subplot(111)
> >
> > cax = ax.imshow(data)
> >
> > ax.set_xticks(range(num//2))
> >
> > fig.show()
> >
> > As a reference, I'm running matplotlib 1.0.0 with python 2.7.0 on a
> > 64-bit Linux (kernel 2.6.31) machine.  I can confirm this on python 2.6
> > and matplotlib 0.99 as well, though.
> >
> > If this isn't expected behavior, I'll be glad to open a bug report.
>
> Joe,
>
> Thanks for the report.  The fix is in svn 8652 and 8653.
>
> Eric
>
> >
> > Thanks!
> > -Joe
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>