torcharrow.ilist_column.ListMethods.vmap

ListMethods.vmap(fun: Callable[[Column], Column])

（实验性 API）矢量化映射。期望一个可调用对象，该对象作用于批次（由 Column 表示）。

示例

>>> import torcharrow as ta
>>> a = ta.column([[1, 2, None, 3], [4, None, 5]])

>>> a
0  [1, 2, None, 3]
1  [4, None, 5]
dtype: List(Int64(nullable=True)), length: 2, null_count: 0

>>> a.list.vmap(lambda col: col + 1)
0  [2, 3, None, 4]
1  [5, None, 6]
dtype: List(Int64(nullable=True), nullable=True), length: 2, null_count: 0

>>> import torcharrow.dtypes as dt
>>> b = ta.column([[(1, "a"), (2, "b")], [(3, "c")]],
    dtype=dt.List(
        dt.Struct([dt.Field("f1", dt.int64), dt.Field("f2", dt.string)])
    ))

>>> b
0  [(1, 'a'), (2, 'b')]
1  [(3, 'c')]
dtype: List(Struct([Field('f1', int64), Field('f2', string)])), length: 2, null_count: 0

>>> b.list.vmap(lambda df: df["f2"])
0  ['a', 'b']
1  ['c']
dtype: List(String(nullable=True), nullable=True), length: 2, null_count: 0