Read any sliceable array into a Dask Dataframe Uses getitem syntax to pull slices out of the array. The array need not be a NumPy array but must support slicing syntax x[50000:100000] and have 2 dimensions: x.ndim == 2 or have a record dtype: x.dtype ==
(arr, chunksize=50_000, columns=None, meta=None)
| 4927 | |
| 4928 | |
| 4929 | def from_array(arr, chunksize=50_000, columns=None, meta=None): |
| 4930 | """Read any sliceable array into a Dask Dataframe |
| 4931 | |
| 4932 | Uses getitem syntax to pull slices out of the array. The array need not be |
| 4933 | a NumPy array but must support slicing syntax |
| 4934 | |
| 4935 | x[50000:100000] |
| 4936 | |
| 4937 | and have 2 dimensions: |
| 4938 | |
| 4939 | x.ndim == 2 |
| 4940 | |
| 4941 | or have a record dtype: |
| 4942 | |
| 4943 | x.dtype == [('name', 'O'), ('balance', 'i8')] |
| 4944 | |
| 4945 | Parameters |
| 4946 | ---------- |
| 4947 | x : array_like |
| 4948 | chunksize : int, optional |
| 4949 | The number of rows per partition to use. |
| 4950 | columns : list or string, optional |
| 4951 | list of column names if DataFrame, single string if Series |
| 4952 | meta : object, optional |
| 4953 | An optional `meta` parameter can be passed for dask |
| 4954 | to specify the concrete dataframe type to use for partitions of |
| 4955 | the Dask dataframe. By default, pandas DataFrame is used. |
| 4956 | |
| 4957 | Returns |
| 4958 | ------- |
| 4959 | dask.DataFrame or dask.Series |
| 4960 | A dask DataFrame/Series |
| 4961 | """ |
| 4962 | import dask.array as da |
| 4963 | |
| 4964 | if isinstance(arr, da.Array): |
| 4965 | return from_dask_array(arr, columns=columns, meta=meta) |
| 4966 | |
| 4967 | from dask.dataframe.dask_expr.io.io import FromArray |
| 4968 | |
| 4969 | result = FromArray( |
| 4970 | arr, |
| 4971 | chunksize=chunksize, |
| 4972 | original_columns=columns, |
| 4973 | meta=meta, |
| 4974 | ) |
| 4975 | if pyarrow_strings_enabled() and arr.dtype.kind in "OU": |
| 4976 | result = expr.ArrowStringConversion(result) |
| 4977 | return new_collection(result) |
| 4978 | |
| 4979 | |
| 4980 | def from_graph(layer, _meta, divisions, keys, name_prefix): |