One of the keys to using stack is to place all of the columns that you do not wish to transform in the index. The dataset in this recipe was initially read with the states in the index. Let's take a look at what would have happened if we did not read the states into the index:

>>> state_fruit2 = pd.read_csv('data/state_fruit2.csv')
>>> state_fruit2

As the state names are not in the index, using stack on this DataFrame reshapes all values into one long Series of values:

>>> state_fruit2.stack()
0  State       Texas
   Apple          12
   Orange         10
   Banana         40
1  State     Arizona
   Apple           9
   Orange          7
   Banana         12
2  State     Florida
   Apple           0
   Orange         14
   Banana        190
dtype: object

This command reshapes all the columns, this time including the states, and is not at all what we need. In order to reshape this data correctly, you will need to put all the non-reshaped columns into the index first with the set_index method, and then use stack. The following code gives a similar result to step 1:

>>> state_fruit2.set_index('State').stack()