pastas.utils.get_equidistant_series

get_equidistant_series(series, freq, minimize_data_loss=False)[source]

Get equidistant timeseries using nearest reindexing.

This method will shift observations to the nearest equidistant timestep to create an equidistant timeseries, if necessary. Each observation is guaranteed to only be used once in the equidistant timeseries.

Parameters
  • series (pandas.Series) – original (non-equidistant) timeseries

  • freq (str) – frequency of the new equidistant timeseries (i.e. “H”, “D”, “7D”, etc.)

  • minimize_data_loss (bool, optional) – if set to True, method will attempt use any unsampled points from original timeseries to fill some remaining NaNs in the new equidistant timeseries. Default is False. This only happens in rare cases.

Returns

s – equidistant timeseries

Return type

pandas.Series

Notes

This method creates an equidistant timeseries with specified freq using nearest sampling (meaning observations can be shifted in time), with additional filling logic that ensures each original measurement is only included once in the new timeseries. Values are filled as close as possible to their original timestamp in the new equidistant timeseries.