get_dummies() 用于多个 Pandas DataFrame

Question

I have a list of DataFrames and I would like to one-hot encode some of the columns in place. For example, if:

In[1]:  df1 = pd.DataFrame(np.array([['a', 'a'], ['b', 'b'], ['c', 'c']]), 
                   columns=['col_1', 'col_2'])
        df2 = pd.DataFrame(np.array([['a', 'a'], ['b', 'b'], ['c', 'c']]),
                   columns=['col_1', 'col_2'])

        combined = [df1, df2]
        combined


Out[1]:    col_1  col_2
        0      a      a
        1      b      b
        2      c      c

I'm currently using the following approach.

In[2]:  for df in combined:
            one_hot = pd.get_dummies(df["col_2"])

            df[one_hot.columns] = one_hot
            df.drop("col_2", axis=1, inplace=True)

        
        df1

Out[2]:      col_1   a   b   c
          0      a   1   0   0
          1      b   0   1   0 
          2      c   0   0   1

Am I missing a more concise solution?

---

Edit

An important requirement is that I need to modify the original dataframes.

Answer 1

I think you can using concat with key which will add a new level of index , then get_dummies

s=pd.concat(combined,keys=range(len(combined)))['col_2'].str.get_dummies()
s['col_1']=pd.concat(combined,keys=range(len(combined)))['col_1'].values

s
Out[20]: 
     a  b  c col_1
0 0  1  0  0     a
  1  0  1  0     b
  2  0  0  1     c
1 0  1  0  0     a
  1  0  1  0     b
  2  0  0  1     c

If you would like to save them into a list for different dfs , you can groupby and save it to dict

d={x:y.reset_index(level=0,drop=True) for x , y in s.groupby(level=0)}
d
Out[16]: 
{0:    a  b  c
 0  1  0  0
 1  0  1  0
 2  0  0  1, 1:    a  b  c
 0  1  0  0
 1  0  1  0
 2  0  0  1}

Answer 2

OP's method is just fine

for df in combined:
    one_hot = pd.get_dummies(df["col_2"])

    df[one_hot.columns] = one_hot
    df.drop("col_2", axis=1, inplace=True)

---

Reassign to all names

df1, df2 = [df.join(pd.get_dummies(df['col_2'])).drop('col_2', 1) for df in combined]