what does "by = lambda x: lambda y: getattr(y, x)" mean?

Question

There is a dataFrame named "subset" and the codes are as follows. pd is the nickname of pandas. I can't figure out the meaning of by = lambda x: lambda y: getattr(y, x).

pivot = pd.pivot_table(subset, values='count', rows=['date'], cols=['sample'], fill_value=0)
by = lambda x: lambda y: getattr(y, x)
grouped = pivot.groupby([by('year'),by('month')]).sum()

Answer 1

by = lambda x: lambda y: getattr(y, x) is equivalent to the following:

def by(x):
    def getter(y):
        return getattr(y, x)
    return getter

getattr(a, b) gets an attribute with the name b from an object named a.

So by('bar') returns a function that returns the attribute 'bar' from an object.

by('bar')(foo) means getattr(foo, 'bar') which is roughly foo.bar.

If that doesn't help, let us know which part you're still having trouble with.