Boolean selection gives much more flexibility than index selection as it is possible to condition on any number of columns. In this recipe, we used a single column as the index. It is possible to concatenate multiple columns together to form an index. For instance, in the following code, we set the index equal to the concatenation of the city and state columns:

>>> college.index = college['CITY'] + ', ' + college['STABBR']
>>> college = college.sort_index()
>>> college.head()

From here, we can select all colleges from a particular city and state combination without boolean indexing. Let's select all colleges from Miami, FL:

>>> college.loc['Miami, FL'].head()

We can compare the speed of this compound index selection with boolean indexing. There is more than an order of magnitude difference:

>>> %%timeit 
>>> crit1 = college['CITY'] == 'Miami' 
>>> crit2 = college['STABBR'] == 'FL'
>>> college[crit1 & crit2]
2.43 ms ± 80.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

>>> %timeit college.loc['Miami, FL']
197 µs ± 8.69 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)