Why don't pandas DataFrame objects support boolean evaluation?

If I have a dataframe, is there some specific reason that direct boolean evaluation isn't supported? Contrary to the pandas error message, it seems absolutely clear and completely unambiguous how to evaluate the truth-value of a dataframe.

What case would a dataframe with shape 0,0 would anyone not want bool(df) -> False? What would that break, or enable, that would be bad?

Conversely, if you have a dataframe with sum(shape) > 0, what would ever not make sense, break, or be enabled by letting bool(df) -> True?

Not asking for philosophy, just like an example or specific reason. To me, it seems very clear how useful this would be, and very opaque that it isn't supported. The error message itself led me to understand the exact opposite of the error message, it almost seems like satire after thinking about it.

In my work the result is, I can't have dataframes flow through a logic system that applies generally, instead i have to make a special case for this single type of data object to use the empty method instead of simply evaluating it in a general condition statement for example.

if a: do_something()
else: do_something_else()

When a is a dataframe, I have to add code to achieve the obvious expected result. It is the only data object I'm working with that requires a separate method to check if it's empty and equate that to being false.

Obviously I can write

if a or (if type(a) is DataFrame and sum(a.shape)>0: do_something

or

if a or not getattr(a, "empty"): do_something

or whatever, tiptoeing around which part of the condition goes where to carefully not bool the dataframe or hardcode the object-specific methods (as I didn't do in these examples).

As written above, the question is what specific harm is avoided by forcing these approaches?

If your answer is to agree with the pandas error message about ambiguity, I guess my question would be what is unambiguous about an all-zero shape being False and any other shape True. Its a fundamental aspect of Python's builtin datatypes that objects with 'shape' are True, and objects that are empty / without shape / zero length are False.

People constantly screw up boolean operations with arrays and dataframes, expecting if checks to broadcast. From what I've seen, this is much more common than people trying to check a dataframe for emptiness. Banning __bool__ helps catch these errors, rather than producing silent misbehavior.

Also, you expect any shape with at least one nonzero dimension to be considered True. That's not what I would expect from an emptiness check, and it's not what DataFrame.empty does. I would expect something that behaves like DataFrame.empty: a dataframe is considered empty if it has at least one zero-length dimension.

Don't make a special case for each if, make a special function for possible conversions

def coerce_to_bool(obj):
    if hasattr(obj, "empty"):
        return obj.empty
    return bool(obj)

    if coerce_to_bool(df):
       ...