Method asflat

[asmeurer]

A method idx.asflat(shape) would return an index j such that a[idx].flatten() == a.flatten()[j].

More generally, we could have a method that reindexes into a reshaped array.

One thing I'm not sure about is how generally this can work. How generally is a flattened index of a simple index still a simple index (i.e., we would not require IntegerIndex, which is not yet implemented)?

[asmeurer]

There are certainly cases where it isn't. For example, if a tuple of slices has some slices with negative steps and others with positive step, then the flattened index would be reshuffled in a way that can't be indexed by a single slice. For example

>>> a = np.arange(100).reshape((10, 10))
>>> a[0:10,10::-1].flatten()
array([ 9,  8,  7,  6,  5,  4,  3,  2,  1,  0, 19, 18, 17, 16, 15, 14, 13,
       12, 11, 10, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 39, 38, 37, 36,
       35, 34, 33, 32, 31, 30, 49, 48, 47, 46, 45, 44, 43, 42, 41, 40, 59,
       58, 57, 56, 55, 54, 53, 52, 51, 50, 69, 68, 67, 66, 65, 64, 63, 62,
       61, 60, 79, 78, 77, 76, 75, 74, 73, 72, 71, 70, 89, 88, 87, 86, 85,
       84, 83, 82, 81, 80, 99, 98, 97, 96, 95, 94, 93, 92, 91, 90])
>>> a.flatten()
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
       17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
       34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50,
       51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67,
       68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84,
       85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99])

At best we could yield pairs of indices (c, i) such that a[idx].flatten()[c] == a.flatten()[c][i].

[asmeurer]

See also https://numpy.org/doc/stable/reference/arrays.indexing.html#flat-iterator-indexing