Hi,
I would like to report two bugs when using twinx with log-scale axes.
<code>
from pylab import *
rcParams['ytick.major.size'] = 12
rcParams['ytick.minor.size'] = 8
ax1 = subplot(2,1,1,
title="Bug twinx + semilogy")
ax1.plot(randn(100)**2,'b+')
ax1.semilogy()
setp(ax1.yaxis.get_minorticklines() + \
ax1.yaxis.get_majorticklines(),
color='b')
ax2 = ax1.twinx()
ax2.plot(randn(100)**0.5,'r.')
ax2.semilogy()
setp(ax2.yaxis.get_minorticklines() + \
ax2.yaxis.get_majorticklines(),
color='r')
ax3 = subplot(2,1,2,
title="Bug twinx + semilogx")
x = arange(100)+1
ax3.plot(x,randn(100)**2,'b+')
ax3.semilogx()
setp(ax3.get_xticklabels(), color='b')
ax4 = ax3.twinx()
ax4.plot(x,randn(100)**0.5,'r.')
setp(ax4.get_xticklabels(), color='r')
show()
</code>
Top: Bug twinx + semilogy.
See the blue and red ticks on the left y-axis: the red ticks should not be there (the right y-axis is correct). Note that the problem does not occur when using linear scales (just comment out the 2 semilogy lines).
Bottom: Bug twinx + semilogx.
The twinx axis (ax4) is linear by default and does not inherit the log scale from the parent axis (ax3).
This is with matplotlib 0.98.3 on Ubuntu 8.10. Maybe these bugs have been corrected in more recent versions...
Cheers.
···
--
.~. Yannick COPIN (o:>* Doctus cum libro
/V\ Institut de physique nucleaire de Lyon (IN2P3-France)
// \\ Lawrence Berkeley National Laboratory
/( )\ aim:YnCopin icq:236931013 http://snovae.in2p3.fr/ycopin/
^`~'^