[SciPy-user] flagging no data in timeseries

Tim Michelsen timmichelsen@gmx-topmail...
Tue Dec 18 14:42:25 CST 2007


Hello,
I while ago I was pointed to the timeseries module from sandbox by one 
of its developers. Thanks again for that. I am still having some 
troubles and questions:

Since I am still very new to the package I need some hints.
In my original data nodata values are marked with "-999".
How can I import the data or create the time series and exclude these no 
data points from further processing?

I presume it has to do with the mask keyword. But I cannot get it.

Here is the code:

data = N.loadtxt(mydata_file, comments='#', delimiter='\t', 
converters=None, skiprows=2, usecols=None, unpack=False)
# define start date
D_hr = TS.Date(freq='HR', year=1995, month=2, day=1, hour=1)
# mask
## prepare time series
# take only needed values
myvalues = data[:,5]
# hourly values
myvalues_ts_hourly = TS.time_series(myvalues, start_date=D_hr)

I will appreciate every help.

For further readability and archivability I write my other questions in 
separate emails.

Kind regards,
Timmie



More information about the SciPy-user mailing list