same code, different outputs between executions (transData weird behaviour)

[avances123]

transData returns different outputs between two executions! this is the code to reproduce it.

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.dates import date2num
import pandas as pd


rng = pd.date_range('01-01-2014','31-12-2014', freq='H')
df = pd.DataFrame(np.random.randn(len(rng),4), index=rng)


def calcula(event):
        ax = event.canvas.figure.gca()
        print ax.transData.transform((date2num(df.index[1]),0))


fig = plt.figure(figsize=(18, 6), dpi=80)
cid = fig.canvas.mpl_connect('draw_event', calcula)
ax = fig.add_subplot(111)
ax.plot_date(x=df.index,y=df[1],fmt="r-")
ax.set_xlim(df.index[1],df.index[111])
ax.grid(True,axis="x")


#plt.show()
print "antes %s" % ax.transData.transform((date2num(df.index[1]),0))
plt.savefig("/tmp/kk.png",dpi=fig.dpi)
print "despues %s" % ax.transData.transform((date2num(df.index[1]),0))

It debugs the transData.transform in several points, the program outputs the same in all points, but if you re-run the program several times, you get different points.

Example in my computer:

---------
antes [ 180.          218.66666667]
[ 180.          218.66666667]
[ 180.          218.66666667]
[ 180.          218.66666667]
despues [ 180.          218.66666667]
---------
antes [ 180.  240.]
[ 180.  240.]
[ 180.  240.]
[ 180.  240.]
despues [ 180.  240.]
----------
antes [ 180.          261.33333333]
[ 180.          261.33333333]
[ 180.          261.33333333]
[ 180.          261.33333333]
despues [ 180.          261.33333333]

[avances123]

my mistake i think, due tu random y data