[SciPy-user] removing certain dates from a time series

Marco Tuckner marcotuckner@public-files...
Thu Apr 3 14:41:22 CDT 2008


Hello,

> > But I don't want to use the data of 29-Feb at all neither masked nor
> > interpoled. If I mask 29-Feb and continue my timeseries operation I would
> > for instance get different monthly or annual averages.
> 
> Please give me an example where masking Feb 29 wouldn't give you the result 
> you expect.
As you are that confident, I will need to investigate further...

> > Therefore I would like to exclude that date completely.

 
> The problem with deleting these dates is that you won't be able to use the 
> convert function afterwards, as it requires a regularly-spaced time series to 
> work.
Ah, OKI I forgot about this.
But then I just would like to bend this date (29-Feb) out when writing the
report [timseries.Report(series)]. How would you do that?
Or would you recommend to reopen the csv-file again afterwards to delete the
records?

Kind regards,
Marco


P.S. I think that many problems I have using timeseries arise from not knowing
how maskedarray works. If only the new maskedarry would have been that well
documented as the time series package [1] I would find my way in much easier!
Is there any documentation for maskedarray like:
[1] http://scipy.org/scipy/scikits/wiki/TimeSeries



More information about the SciPy-user mailing list