I'm trying to wrap my brain around pandas data structures and trying to use them in anger a bit. I've figured out that groupby operations result in a pandas series object. But I can't quite figure out how to use the resulting series. In particular, I want to do two thing:
1) "join" the results back to the initial DataFrame
2) select a specific value from the resulting series based on the hierarchical index.
Here's a toy example to work with:
import pandas
df = pandas.DataFrame({'group1': ['a','a','a','b','b','b'],
'group2': ['c','c','d','d','d','e'],
'value1': [1.1,2,3,4,5,6],
'value2': [7.1,8,9,10,11,12]
})
dfGrouped = df.groupby( ["group1", "group2"] , sort=True)
## toy function, obviously not my real function
def fun(x): return mean(x**2)
results = dfGrouped.apply(lambda x: fun(x.value1))
so the resulting series (results) looks like this:
group1 group2
a c 2.605
d 9.000
b d 20.500
e 36.000
That makes sense. But how do I:
1) join this back to the original DataFrame df
2) Select a single value where, say, group1=='b' & group2=='d'