使用三列的分组问题创建一个数据框

Question

I have the following dataframe:

       A               B                  C
  I am motivated     Agree                4
  I am motivated     Strongly Agree       5
  I am motivated     Disagree             6
  I am open-minded   Agree                4
  I am open-minded   Disagree             4
  I am open-minded   Strongly Disagree    3

Where column A is the question, column B is the answer, and column C is the frequency of "Strongly Agree", "Agree", "Disagree", and "Strongly Disagree" for the questions in column A.

How can I convert it into the following dataframe?

                  Strongly Agree    Agree     Disagree   Strongly Disagree
I am motivated        5               4           6             0
I am open-minded      0               4           4             3

I tried looking at groupby() for columns from other posts but could not figure it out. Using python 3

Answer 1

Use DataFrame.pivot_table() method:

In [250]: df.pivot_table(index='A', columns='B', values='C', aggfunc='sum', fill_value=0)
Out[250]:
B                 Agree  Disagree  Strongly Agree  Strongly Disagree
A
I am motivated        4         6               5                  0
I am open-minded      4         4               0                  3

Answer 2

Because these are already frequency counts, we can assume that we have unique Question / Opinion pairs. So, we can use set_index and unstack as there won't be a need to aggregate. This should save us some time with efficiency. We could accomplish the same goal with pivot , however, pivot doesn't have a fill_value option that enables us to preserve dtype

df.set_index(['A', 'B']).C.unstack(fill_value=0)

B                 Agree  Disagree  Strongly Agree  Strongly Disagree
A                                                                   
I am motivated        4         6               5                  0
I am open-minded      4         4               0                  3

---

Extra Credit
Turn 'B' into a pd.Categorical and the columns will be sorted

df.B = pd.Categorical(
    df.B, ['Strongly Disagree', 'Disagree', 'Agree', 'Strongly Agree'], True)
df.set_index(['A', 'B']).C.unstack(fill_value=0)

B                 Strongly Disagree  Disagree  Agree  Strongly Agree
A                                                                   
I am motivated                    0         6      4               5
I am open-minded                  3         4      4               0