Question

When I create a DataFrame, then sort by a column it appears to be sorted in the iteractive display (i.e. whatever repr gives), but when I call the DataFrame.plot() function it plots the unsorted array. Calling matplotlib.pylab.plot works fine though. I suspect it’s something to do with clever pointer rearrangement not being passed to whatever the plot function is calling to access the data … or maybe I’m just doing something dumb. I’ve tried this on pandas 0.8.1 (osx and linux with python2.7.something) and pandas 0.9.0 (osx with python3.something).

import pandas
import numpy
from matplotlib.pylab import *
a = numpy.random.randn(100,10)
df = pandas.DataFrame(a)
df.shape
df.sort(column=0)
df.columns
df.sort(column=0, inplace=True)
df[0]
df[0].plot()

Answer 1

Plotting is by default your data versus the index. In case of a numerical index, look at the index after sorting on a column. you might want to use plot(use_index=False). See example below.

In [2]: df = pd.DataFrame(np.random.randn(10, 2), columns=['A', 'B'])

In [3]: df
Out[3]:
          A         B
0 -0.938196  2.220319
1 -0.022503  0.564602
2  0.033094 -0.717969
3  2.466486  1.229651
4 -0.641596 -1.016921
5  0.094125  1.531526
6  0.579631  1.398635
7 -0.854799 -0.930904
8 -1.177894 -1.501657
9  0.341655 -0.917243

In [4]: df.sort(columns='A', inplace=True)
Out[4]:
          A         B
8 -1.177894 -1.501657
0 -0.938196  2.220319
7 -0.854799 -0.930904
4 -0.641596 -1.016921
1 -0.022503  0.564602
2  0.033094 -0.717969
5  0.094125  1.531526
9  0.341655 -0.917243
6  0.579631  1.398635
3  2.466486  1.229651

In [5]: df['A'].plot(use_index=False)
Out[5]: <matplotlib.axes.AxesSubplot at 0xb56ac6c>