QTable equality broken with numpy 1.23

Description

When updating from numpy 1.22 to numpy 1.23, trying to test for equality with an empty and non-empty table rasies an error

Expected behavior

False to be printed without error. With numpy 1.22 this is the output:

/Users/dstansby/mambaforge/envs/sunpy/lib/python3.10/site-packages/astropy/table/table.py:3474: FutureWarning: elementwise == comparison failed and returning scalar instead; this will raise an error or perform elementwise comparison in the future.
  result = self.as_array() == other
False

Actual behavior

With numpy 1.23:

Traceback (most recent call last):
  File "/Users/dstansby/github/sunpy/test.py", line 6, in <module>
    print(t1 == t2)
  File "/Users/dstansby/mambaforge/envs/sunpy/lib/python3.10/site-packages/astropy/table/table.py", line 3426, in __eq__
    return self._rows_equal(other)
  File "/Users/dstansby/mambaforge/envs/sunpy/lib/python3.10/site-packages/astropy/table/table.py", line 3474, in _rows_equal
    result = self.as_array() == other
TypeError: Cannot compare structured arrays unless they have a common dtype.  I.e. `np.result_type(arr1, arr2)` must be defined.

Steps to Reproduce

from astropy.table import QTable
import numpy as np

t1 = QTable([np.array([], dtype=float)], names=('a'))
t2 = QTable([np.array([1], dtype=float)], names=('b'))
print(t1 == t2)

System Details

macOS-12.4-arm64-arm-64bit Python 3.10.4 | packaged by conda-forge | (main, Mar 24 2022, 17:42:03) [Clang 12.0.1 ] Numpy 1.23.0 pyerfa 2.0.0.1 astropy 5.1 Scipy 1.8.1 Matplotlib 3.5.2

[mhvk]

Indeed, it is the try part that goes from being cheap to being essentially free, so it pays to put the common path inside a try/except and have all the costs born by the uncommon path (or error).

[taldcroft]

Oh, really? I was told that “except” is expensive.

Yes, if the except path is taken (exception gets raised) that is expensive, but the try is cheap.