Do structural, historical or philosophical reason(s) explain why native python allow slicing list of lists only with list comprehension? [closed]

Using a list comprehension is the straightforward way.

table[0:2][1:4] doesn't do what you want because table[0:2] is a list of the first two elements of table, so [1:4] takes a slice of that list which still just contains elements of table. Both are slice operations on two-dimensional lists, so there is no semantic reason that one slice should return some unchanged rows while the other slice should return all rows but with only some of the columns.

The reason Python does not include such an operation on the built-in list type is probably that the comprehensions are easy enough to write, simple enough to read to understand their meaning, and the need for multi-dimensional slicing of native lists is not that common.

Note also that these are two different kinds of "multi-dimensional slice" - one returns a single column as a one-dimensional list the other returns a new two-dimensional list. Whether you want a sliced column to be one- or two-dimensional will depend on the use-case, and if the built-in list class took a stance on which one is "correct" then it could easily lead to problems for some users who expect the other behaviour.

If you need multi-dimensional matrices of elements, the numpy package supports a clean syntax for multi-dimensional slicing. You may already be using it for multi-dimensional data anyway, and if you're not then you should consider it.

>>> import numpy
>>> m = numpy.array([list(range(10)), list(range(10)), list(range(10))])
>>> m[:, 0]
array([0, 0, 0])
>>> m[0:2, 1:4]
array([[1, 2, 3],
       [1, 2, 3]])