BUG: plotting daily data with two-day gaps

[lesteve]

Hi,

you should be able to copy and paste this in a ipython --pylab session:

import pandas
index = pandas.date_range(start='2001-1-1', end='2001-1-10', freq='D')
data = index.day
ts = pandas.TimeSeries(index=index, data=data)
tsWithGaps = ts[:3].append(ts[5:])
figure()
tsWithGaps.plot(style='o')
figure()
tsWithGaps.asfreq('D').plot(style='o')

I tried this snippet with 0.8.2.dev-7c54141 (i.e. latest git version as I write). The two plots below come from an improved version of the snippet which prints the timeseries on the plot.

The first plot I get looks weird (wrong label for January 5th + double data point on the 8th)
http://imgur.com/LFJup,aYLYJ#0
The timeseries is printed on the plot and looks perfectly fine.

The second plot with the explicitly resampled ts is fine:
http://imgur.com/LFJup,aYLYJ#1

Trying to tweak the parameters a 3-day gap looks fine and you get x axis labelling like plt.plot_date. She 2-day gaps seems always broken, although in a more subtle way than the plot above (generally the label before the gap is wrong but I don't get double data points). It looks like a timeseries with 2-day gaps is close enough to a regular index timeseries (maybe because of week-ends) and the index labelling goes wrong somewhere.

[changhiskhan]

Just put in a fix for this. Please give it a shot and let me know if it works on your end.
Thanks for the bug report

[lesteve]

Just tried and it looks fine. Thanks a lot for the fix. I guess I can still explicitly resample the timeseries if I want to get the nicer scikits.timeseries style tick labeling.