Understanding matplotlib.subplots python

Question

I built a set of pie charts with some help. Insert an image into a slice of a pie chart. My charts look great, now I need to put all 6 of them in a 2x3 shape with common marks on the common x and y axis. To start, I look at the subtitles and thought that I could make it work. I downloaded a few examples and started trying a few things.

f, (a) = (plt.subplots(nrows=1, ncols=1, sharex=True, sharey=True))#, #squeeze=False, subplot_kw=None, gridspec_kw=None)) print(type(f),'\n',type(a),'\n')#,type(b)) 

gives:

class 'matplotlib.figure.Figure'

class 'matplotlib.axes._subplots.AxesSubplot'

a

  f, (a) = (plt.subplots(nrows=1, ncols=1, sharex=True, sharey=True, squeeze=False, subplot_kw=None, gridspec_kw=None)) print(type(f),'\n',type(a),'\n')#,type(b)) 

returns:

class 'matplotlib.figure.Figure'

class 'numpy.ndarray'

When I do this:

 f, (a,b) = (plt.subplots(nrows=2, ncols=1, sharex=True, sharey=True, squeeze=False, subplot_kw=None, gridspec_kw=None)) print(type(f),'\n',type(a),'\n',type(b)) 

I get similar results, however if nrows = 1 and ncols = 2, I get an error message:

  f, (a,b) = (plt.subplots(nrows=1, ncols=2, sharex=True, sharey=True, squeeze=False, subplot_kw=None, gridspec_kw=None)) print(type(f),'\n',type(a),'\n',type(b)) 

ValueError: not enough values ​​to unpack (expected 2, received 1)

but again:

  f, (a , b) = ( plt.subplots(nrows=1, ncols=2, sharex=True, sharey=True))#, #squeeze=False, subplot_kw=None, gridspec_kw=None)) print(type(f),'\n',type(a),'\n',type(b)) 

gives class 'matplotlib.figure.Figure'

class 'matplotlib.axes._subplots.AxesSubplot'

class 'matplotlib.axes._subplots.AxesSubplot'

Why is it either an array or an axis, and also why is there no work of 2X1 and 1X2? I want from heaven, I could better understand the documentation. Thanks.

Answer 1

The various types of data returned are associated with the squeeze keyword plt.subplots() argument, which is set to True by default. Let me improve the documentation with the corresponding unpacks:

compress: bool, optional, default: True

• If True, additional dimensions are extruded from the returned Axes object:

  • if only one subplot is built (nrows = ncols = 1), the resulting single Axes object is returned as a scalar.
    fig, ax = plt.subplots()
  • for Nx1 or 1xN subnets, the returned object is an array of 1D numpy objects of Axes objects that are returned as numpy 1D arrays.
    fig, (ax1, ..., axN) = plt.subplots(nrows=N, ncols=1) (for Nx1)
    fig, (ax1, ..., axN) = plt.subplots(nrows=1, ncols=N) (for 1xN)
  • for NxM, subsets with N> 1 and M> 1 are returned as 2D arrays.
    fig, ((ax11, .., ax1M),..,(axN1, .., axNM)) = plt.subplots(nrows=N, ncols=M)
• If False, no compression is performed at all: the returned Axes object is always a 2D array containing Axes instances, even if it ends with 1x1.
  fig, ((ax,),) = plt.subplots(nrows=1, ncols=1, squeeze=False)
fig, ((ax,), .. ,(axN,)) = plt.subplots(nrows=N, ncols=1, squeeze=False) for Nx1
  fig, ((ax, .. ,axN),) = plt.subplots(nrows=1, ncols=N, squeeze=False) for 1xN
  fig, ((ax11, .., ax1M),..,(axN1, .., axNM)) = plt.subplots(nrows=N, ncols=M)

Alternatively, you can always use the unpacked version.

 fig, ax_arr = plt.subplots(nrows=N, ncols=M, squeeze=False) 

and index the array to get the axes, ax_arr[1,2].plot(..) .

So, for a 2 x 3 grid, it really doesn't matter if you set squeeze to False . The result will always be a two-dimensional array. You can unpack it as

 fig, ((ax1, ax2, ax3),(ax4, ax5, ax6)) = plt.subplots(nrows=2, ncols=3) 

have ax{i} as matplotlib axis objects, or you can use the packaged version

 fig, ax_arr = plt.subplots(nrows=2, ncols=3) ax_arr[0,0].plot(..) # plot to first top left axes ax_arr[1,2].plot(..) # plot to last bottom right axes