bach.DataFrame.fillna

fillna​

(*, value=None, method=None, axis=0, sort_by=None, ascending=True, window=None)

​[source]

Fill any NULL value using a method or with a given value.

Parameters​

  • value (Union[Series, int, numpy.int64, float, numpy.float64, bool, str, None, datetime.date, datetime.time, datetime.datetime, numpy.datetime64, datetime.timedelta, numpy.timedelta64, uuid.UUID, pandas._libs.interval.Interval, dict, list, Dict[str, Union[int, numpy.int64, float, numpy.float64, bool, str, None, datetime.date, datetime.time, datetime.datetime, numpy.datetime64, datetime.timedelta, numpy.timedelta64, uuid.UUID, pandas._libs.interval.Interval, dict, list, Series]]]) – A literal/series to fill all NULL values on each series or a dictionary specifying which literal/series to use for each series.

  • method (Optional[str]) – Method to use for filling NULL values on all DataFrame series. Supported values:

    • ”ffill”/”pad”: Fill missing values by propagating the last non-nullable value in the series.
    • ”bfill”/”backfill”: Fill missing values with the next non-nullable value in the series.

  • axis (int) – only axis=0 is supported.

  • sort_by (Optional[Union[str, Sequence[str]]]) – series label or sequence of labels used to sort values. Sorting of values is needed since result might be non-deterministic, as rows with NULLs might yield different results affecting the values to be propagated when using a filling method.

  • ascending (Union[bool, List[bool]]) – Whether to sort ascending (True) or descending (False). If this is a list, then the sort_by must also be a list and len(ascending) == len(sort_by).

  • window (Optional[Union[Window, DataFrame]]) – If provided, values are propagated within each partitioning group only, otherwise values are propagated in the entire DataFrame. Considered only when method param is provided.

Returns​

a new dataframe with filled missing values.

Return type​

DataFrame

note

sort_by is required if method is specified and the DataFrame has no order_by.

danger

If sort_by is non-deterministic, this operation might yield different results after performing other operations over the resultant dataframe.