根据答案https://stackoverflow.com/a/5347492/7505395改编为Interweaving two numpy arrays:
We should not interleave numpy‘s vectorized operation along with a Python loop. This slows down performance drastically, as the code is iterating using native Python. 我们不应该将 numpy的向量化操作与Python循环一起插入。 由于代码正在使用本机Python进行迭代,因此这会极大地降低性能。 For example, the...
which generate theJaggedArrays. This is an illustration of how each Awkward class provides one feature, and you get desired behavior by combining them.
(Note that a UnionArray may be the best way to interleave two arrays, even if they have the same type. Heterogeneity is not a necessary feature of a UnionArray.) Numpy has a dtype=object for arrays of Python objects, but awkward's ObjectArray creates Python objects on demand from array...