OOM when re-binning binned data array in same dimensions and using .bin(..., dim=da.dims) #3872

To reproduce

import scipp as sc
# Binned data array with shape (4000, 4000)
da = sc.data.binned_xy(100_000, 4000, 4000)
da.bin(x=100, y=100, dim=da.dims)

Output

Killed

Expected result
A binned data array with two dimensions (x, y) each having 100 bins.

Note that

import scipp as sc
da = sc.data.binned_xy(100_000, 4000, 4000)
da.bin(x=100, z=100, dim=da.dims)

runs and produces a binned data array with two dimensions (x, z) each having 100 bins.

So the issue seems to have to do with the dimension names of the original data array being the same as the requested binning dimensions.