def to_pandas(self, index: pd.Index = None, **kwargs) -> pd.Series:
if self.categories.dtype.kind == "f":
new_mask = bools_to_mask(self.notnull())
col = column.build_categorical_column(
categories=self.categories,
codes=column.as_column(self.codes, dtype=self.codes.dtype),
mask=new_mask,
ordered=self.dtype.ordered,
size=self.codes.size,
)
else:
col = self
signed_dtype = min_signed_type(len(col.categories))
codes = col.cat().codes.astype(signed_dtype).fillna(-1).to_array()
categories = col.categories.dropna(drop_nan=True).to_pandas()
data = pd.Categorical.from_codes(codes,
categories=categories,
ordered=col.ordered)
return pd.Series(data, index=index)
def pandas_categorical_as_column(categorical, codes=None):
"""Creates a CategoricalColumn from a pandas.Categorical
If ``codes`` is defined, use it instead of ``categorical.codes``
"""
codes = categorical.codes if codes is None else codes
codes = column.as_column(codes)
valid_codes = codes != codes.dtype.type(-1)
mask = None
if not valid_codes.all():
mask = bools_to_mask(valid_codes)
return column.build_categorical_column(
categories=categorical.categories,
codes=column.as_column(codes.base_data, dtype=codes.dtype),
size=codes.size,
mask=mask,
ordered=categorical.ordered,
)
def timeseries(
start="2000-01-01",
end="2000-01-31",
freq="1s",
dtypes=None,
nulls_frequency=0,
seed=None,
):
"""Create timeseries dataframe with random data
Parameters
----------
start : datetime (or datetime-like string)
Start of time series
end : datetime (or datetime-like string)
End of time series
dtypes : dict
Mapping of column names to types.
Valid types include {float, int, str, 'category'}.
If none is provided, this defaults to
``{"name": "category", "id": int, "x": float, "y": float}``
freq : string
String like '2s' or '1H' or '12W' for the time series frequency
nulls_frequency : float
Fill the series with the specified proportion of nulls. Default is 0.
seed : int (optional)
Randomstate seed
Examples
--------
>>> import cudf as gd
>>> gdf = gd.datasets.timeseries()
>>> gdf.head() # doctest: +SKIP
timestamp id name x y
2000-01-01 00:00:00 967 Jerry -0.031348 -0.040633
2000-01-01 00:00:01 1066 Michael -0.262136 0.307107
2000-01-01 00:00:02 988 Wendy -0.526331 0.128641
2000-01-01 00:00:03 1016 Yvonne 0.620456 0.767270
2000-01-01 00:00:04 998 Ursula 0.684902 -0.463278
"""
if dtypes is None:
dtypes = {"name": "category", "id": int, "x": float, "y": float}
index = pd.DatetimeIndex(
pd.date_range(start, end, freq=freq, name="timestamp"))
state = np.random.RandomState(seed)
columns = {k: make[dt](len(index), state) for k, dt in dtypes.items()}
df = pd.DataFrame(columns, index=index, columns=sorted(columns))
if df.index[-1] == end:
df = df.iloc[:-1]
gdf = cudf.from_pandas(df)
for col in gdf:
mask = state.choice(
[True, False],
size=len(index),
p=[1 - nulls_frequency, nulls_frequency],
)
mask_buf = bools_to_mask(cudf.core.column.as_column(mask))
masked_col = gdf[col]._column.set_mask(mask_buf)
gdf[col] = cudf.Series._from_data(ColumnAccessor({None: masked_col}),
index=gdf.index)
return gdf
