How does Pandas resolve the function specified by name in df.agg

问题

I'm learning Pandas and Numpy, currently going through this section of the tutorial. I'm new to Python altogether, so this is probably a basic beginner's question.

Given this data frame:

df = pd.DataFrame(np.random.randn(4, 3), columns=['A', 'B', 'C'],
                      index=pd.date_range('1/1/2000', periods=4))
df.iloc[3:7] = np.nan

I can't explain the difference between the following results of df.agg:

Call 1:

df.agg(sum)

#Result:
A   NaN
B   NaN
C   NaN
dtype: float64

Call 2:

df.agg('sum')

#Result:
A   -1.776752
B   -2.070156
C   -0.124162
dtype: float64

The result of df.agg('sum') is the same as that of df.agg(np.sum) or df.sum(). I expected df.agg('sum') to produce the same result as df.agg(sum).

Does Pandas have special logic to resolve these functions such that it would prefer np.sum (or run df.sum) instead of the built-in sum?

回答1:

In the documentation you linked to, it says:

You can also pass named methods as strings.

So strings are resolved as method names on the DataFrame (or Series, if you call agg on a Series).

来源：https://stackoverflow.com/questions/53436538/how-does-pandas-resolve-the-function-specified-by-name-in-df-agg