scmdata.timeseries

TimeSeries handling

Functionality for handling and storing individual time-series

TimeSeries

class TimeSeries(data, time=None, **kwargs)[source]

Bases: OpsMixin

A 1D time-series with metadata

Proxies an xarray.DataArray with a single time dimension

Create a deep copy of the timeseries.

Any further modifications to the Timeseries returned copy will not be reflected in the current Timeseries

interpolate(target_times, interpolation_type='linear', extrapolation_type='linear')[source]

Interpolate the timeseries onto a new time axis

Parameters:

target_times (typing.Union[numpy.ndarray, typing.List[typing.Union[datetime.datetime, int]]]) – Time grid onto which to interpolate
interpolation_type (str) – Interpolation type. Options are ‘linear’
extrapolation_type (str or None) – Extrapolation type. Options are None, ‘linear’ or ‘constant’

Returns:

TimeSeries – A new TimeSeries with the new time dimension

property meta

Metadata associated with the timeseries

property name

Timeseries name

If no name was provided this will be an automatically incrementing number

reindex(time, **kwargs)[source]

Update the time dimension, filling in the missing values with NaN’s

This is different to interpolating to fill in the missing values. Uses xarray.DataArray.reindex to perform the reindexing

Parameters:

time (obj:np.ndarray) – Time values to reindex the data to. Should be np.datetime64 values
**kwargs (typing.Any) – Additional arguments passed to xarray’s DataArray.reindex function

Returns:

TimeSeries – A new TimeSeries with the new time dimension

References

property time_points

Time points of the data

property values

Get the data as a numpy array