Histogram bug introduced in r8218

It seems that changes introduced in r8218 drastically changed how `hist` handles Python lists.

For example, the histogram given by the following snippet, works as expected:

x = np.random.randn(100)
plt.hist(x)

However, if you pass a 1D list to `hist`, the 1D list is cast to a list of length-1 arrays---each treated as a separate histogram. Continuing the above code, check the results of the following:

plt.hist(list(x))

I assume that this behavior is unintended, right?

-Tony

It seems that changes introduced in r8218 drastically changed how `hist` handles Python lists.

For example, the histogram given by the following snippet, works as expected:

x = np.random.randn(100)
plt.hist(x)

However, if you pass a 1D list to `hist`, the 1D list is cast to a list of length-1 arrays---each treated as a separate histogram. Continuing the above code, check the results of the following:

plt.hist(list(x))

I assume that this behavior is unintended, right?

Fixed in svn 8293.

Eric

ยทยทยท

On 05/03/2010 01:53 PM, Tony S Yu wrote:

-Tony
------------------------------------------------------------------------------
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
matplotlib-devel List Signup and Options