Function _nonempty_index

dask/dataframe/backends.py:380–451

(idx)

@meta_nonempty.register(pd.Index)
def _nonempty_index(idx):
    typ = type(idx)
    if typ is pd.RangeIndex:
        return pd.RangeIndex(2, name=idx.name, dtype=idx.dtype)
    elif is_any_real_numeric_dtype(idx):
        return typ([1, 2], name=idx.name, dtype=idx.dtype)
    elif typ is pd.DatetimeIndex:
        start = "1970-01-01"
        # Need a non-monotonic decreasing index to avoid issues with
        # partial string indexing see https://github.com/dask/dask/issues/2389
        # and https://github.com/pandas-dev/pandas/issues/16515
        # This doesn't mean `_meta_nonempty` should ever rely on
        # `self.monotonic_increasing` or `self.monotonic_decreasing`
        try:
            return pd.date_range(
                start=start,
                periods=2,
                freq=idx.freq,
                tz=idx.tz,
                name=idx.name,
                unit=idx.unit,
            )
        except ValueError:  # older pandas versions
            data = [start, "1970-01-02"] if idx.freq is None else None
            return pd.DatetimeIndex(
                data, start=start, periods=2, freq=idx.freq, tz=idx.tz, name=idx.name
            )
    elif typ is pd.PeriodIndex:
        return pd.period_range(
            start="1970-01-01", periods=2, freq=idx.freq, name=idx.name
        )
    elif typ is pd.TimedeltaIndex:
        start = np.timedelta64(1, "D")
        try:
            return pd.timedelta_range(
                start=start, periods=2, freq=idx.freq, name=idx.name
            )
        except ValueError:  # older pandas versions
            start = np.timedelta64(1, "D")
            data = [start, start + 1] if idx.freq is None else None
            return pd.TimedeltaIndex(
                data, start=start, periods=2, freq=idx.freq, name=idx.name
            )
    elif typ is pd.CategoricalIndex:
        if len(idx.categories) == 0:
            data = pd.Categorical(_nonempty_index(idx.categories), ordered=idx.ordered)
        else:
            data = pd.Categorical.from_codes(
                [-1, 0], categories=idx.categories, ordered=idx.ordered
            )
        return pd.CategoricalIndex(data, name=idx.name)
    elif typ is pd.MultiIndex:
        levels = [_nonempty_index(l) for l in idx.levels]
        codes = [[0, 0] for i in idx.levels]
        try:
            return pd.MultiIndex(levels=levels, codes=codes, names=idx.names)
        except TypeError:  # older pandas versions
            return pd.MultiIndex(levels=levels, labels=codes, names=idx.names)