python - Pandas groupby 用于一列中的多个值

Question

我有一个类似于下面的数据框

+----------------+-------+
| class          | year  |
+----------------+-------+
| ['A', 'B']     | 2001  |
| ['A']          | 2002  |
| ['B']          | 2001  |
| ['A', 'B', 'C']| 2003  |
| ['B', 'C']     | 2001  |
| ['C']          | 2003  |
+----------------+-------+

我想使用它创建一个数据框，以便生成的表格显示每年类(class)中每个类别的计数。

+-----+----+----+----+
|year | A  | B  | C  |
+-----+----+----+----+
|2001 | 1  | 3  | 1  |
|2002 | 1  | 0  | 0  |
|2003 | 1  | 1  | 2  |
+-----+----+----+----+

最简单的方法是什么？

Answer 1

尝试 unnesting

s=unnesting(df,['class'])

然后，我们做crosstab

pd.crosstab(s['year'],s['class'])

方法来自sklearn

from sklearn.preprocessing import MultiLabelBinarizer
mlb = MultiLabelBinarizer()
pd.DataFrame(mlb.fit_transform(df['class']),columns=mlb.classes_, index=df.year).sum(level=0)
Out[293]: 
      A  B  C
year         
2001  2  2  1
2002  1  1  1
2003  0  1  1

---

get_dummies的方法

df.set_index('year')['class'].apply(','.join).str.get_dummies(sep=',').sum(level=0)
Out[297]: 
      A  B  C
year         
2001  2  2  1
2002  1  1  1
2003  0  1  1

---

def unnesting(df, explode):
    idx = df.index.repeat(df[explode[0]].str.len())
    df1 = pd.concat([
        pd.DataFrame({x: np.concatenate(df[x].values)}) for x in explode], axis=1)
    df1.index = idx

    return df1.join(df.drop(explode, 1), how='left')