[SciPy-user] Still having plotting issue with latest svnscikits.timeseries

Pierre GM pgmdevlist@gmail....
Wed Dec 10 13:32:02 CST 2008


On Dec 10, 2008, at 1:55 PM, Dharhas Pothina wrote:

> Hi Pierre,
>
> The culprit seems to be the .fill_missing_dates() method. If I  
> comment that out my plots are fine and if I use it I have the  
> problems I mentioned.

Should have thought about it earlier. When you use .fill_missing_dates  
on your data, you introduce a lot of missing values. matplotlib  
doesn't know how to connect those missing values with lines, so it  
doesn't plot the lines. However, it plots the dots alright. The pb is  
thus a limitation of matplotlib, not of timeseries (relief), and no,  
there's no work around.

>
>> Since I can plot after converting using .asfreq this isn't too much
>
> Disregard my comment about .asfreq working I'm unable to reproduce  
> it now. .convert only works when converting to a lower frequency.

convert transforms a series w/ one frequency to another series w/  
another frequency. From lower to higher frequency, it outputs by  
default a 2D array, which is not what you want. Check the docstring.
asfreq just convert the dates from one frequency to another. The data  
is left unchanged. That's what you want. Here again, please refer to  
the documentation.

>
>> OK, thx for the dataset, but please, don't copy-paste directly from
>> ipython. Just send us the commands, without a prompt (the In[...]),  
>> in
>> a script if you prefer. Right now, I don't have time to spend on
>
> Sorry about that. I've attached a script called test.py and the data  
> JDM3_short.txt

OK, thx a lot. So, yes, that was only a pb of masked data.


More information about the SciPy-user mailing list