-
+ 
-
+
from __future__ import annotations
@@ -386,7 +409,7 @@ Source code for sharrow.dataset
[docs]def construct(source):
"""
- A generic constructor for creating Datasets from various similar objects.
+ Create Datasets from various similar objects.
Parameters
----------
@@ -418,7 +441,7 @@ Source code for sharrow.dataset
sparse: bool = False,
preserve_cat: bool = True,
) -> Dataset:
- """Convert a pandas.DataFrame into an xarray.Dataset
+ """Convert a pandas.DataFrame into an xarray.Dataset.
Each column will be converted into an independent variable in the
Dataset. If the dataframe's index is a MultiIndex, it will be expanded
@@ -436,6 +459,13 @@ Source code for sharrow.dataset
If true, create a sparse arrays instead of dense numpy arrays. This
can potentially save a large amount of memory if the DataFrame has
a MultiIndex. Requires the sparse package (sparse.pydata.org).
+ preserve_cat : bool, default True
+ If true, preserve encoding of categorical columns. Xarray lacks an
+ official implementation of a categorical datatype, so sharrow's
+ dictionary-based digital encoding is applied instead. Note that in
+ native xarray usage, the resulting variable will look like integer
+ values instead of the category values. The `dataset.cat` accessor
+ can be used to interact with the categorical data.
Returns
-------
@@ -446,7 +476,6 @@ Source code for sharrow.dataset
xarray.DataArray.from_series
pandas.DataFrame.to_xarray
"""
-
# this is much faster than the default xarray version when not
# using a MultiIndex.
@@ -515,7 +544,7 @@ Source code for sharrow.dataset
index=None,
):
"""
- Convert a pyarrow.Table into an xarray.Dataset
+ Convert a pyarrow.Table into an xarray.Dataset.
Parameters
----------
@@ -620,7 +649,6 @@ Source code for sharrow.dataset
-------
Dataset
"""
-
# handle both larch.OMX and openmatrix.open_file versions
if "lar" in type(omx).__module__:
omx_data = omx.data
@@ -995,9 +1023,7 @@ Source code for sharrow.dataset
@xr.register_dataset_accessor("single_dim")
class _SingleDim:
- """
- Convenience accessor for single-dimension datasets.
- """
+ """Convenience accessor for single-dimension datasets."""
__slots__ = ("dataset", "dim_name")
@@ -1139,9 +1165,7 @@ Source code for sharrow.dataset
@xr.register_dataarray_accessor("single_dim")
class _SingleDimArray:
- """
- Convenience accessor for single-dimension datasets.
- """
+ """Convenience accessor for single-dimension datasets."""
__slots__ = ("dataarray", "dim_name")
@@ -1494,7 +1518,7 @@ Source code for sharrow.dataset
from .relationships import sparse_array_type
def to_numpy(var):
- """Coerces wrapped data to numpy and returns a numpy.ndarray"""
+ """Coerces wrapped data to numpy and returns a numpy.ndarray."""
data = var.data
if hasattr(data, "chunks"):
data = data.compute()
@@ -1518,7 +1542,7 @@ Source code for sharrow.dataset
@register_dataset_method
def select_and_rename(self, name_dict=None, **names):
"""
- Select and rename variables from this Dataset
+ Select and rename variables from this Dataset.
Parameters
----------
@@ -1682,22 +1706,20 @@ Source code for sharrow.dataset
-
+
+