Semantics between Series.apply and Series.agg incorrect

[mhriemers]

Correct me if I'm wrong, but isn't Series.agg supposed to reduce all rows to a singular value and Series.apply supposed to apply an operation to each individual row? Currently, when a function doesn't throw one of ValueError, AttributeError, TypeError, it is assumed to be a function meant for Series.apply even when passed to Series.agg.

Consider the following example:

def one_sample_t_test(r, mu_0):
    return (np.mean(r) - mu_0) / (np.std(r) / np.sqrt(len(r)))

s.agg(lambda r: one_sample_t_test(r, 0))

This generates exactly the same series as before because the function is called with individual values instead of the whole frame. When used in conjunction with another aggregator, like mean, it crashes since it can't concatenate a singular value with a series.

pandas/pandas/core/apply.py

Lines 1040 to 1051 in c293caf

	# try a regular apply, this evaluates lambdas
	# row-by-row; however if the lambda is expected a Series
	# expression, e.g.: lambda x: x-x.quantile(0.25)
	# this will fail, so we can try a vectorized evaluation
	# we cannot FIRST try the vectorized evaluation, because
	# then .agg and .apply would have different semantics if the
	# operation is actually defined on the Series, e.g. str
	try:
	result = self.obj.apply(f)
	except (ValueError, AttributeError, TypeError):
	result = f(self.obj)

Why would somebody ever call Series.agg with a function that only works on singular values?

[mhriemers]

Related to #49673.

[rhshadrach]

Thanks for the report - I think this is a duplicate of #49673. Please comment here if you think this is not accurate and can reopen.