用 3 列重塑 pandas 数据框答案

【问题标题】：Reshape pandas dataframe with 3 columns用 3 列重塑 pandas 数据框
【发布时间】：2021-07-08 16:31:07
【问题描述】：

我有以下代码：

sentiments = ['He is good', 'He is bad', 'She love her', 'She is fine with it', 'I like going outside', 'Its okay']
positive = [1,0,1,0,1,0]
negative = [0,1,0,0,0,0]
neutral = [0,0,0,1,0,1]
neutral
df = pd.DataFrame({'Sentiments':sentiments, 'Positives':positive, 'Negatives': negative, 'Neutrals':neutral})
df.head()

创建这个：

我希望只有 2 列，1 列带有情绪，其他带有类别，这应该是特定的情绪，即结果应该是：

Sentiment	Category
abc	positive
xmy	negative
poi	neutral

【问题讨论】：

标签： python pandas dataframe nlp sentiment-analysis

【解决方案1】：

试试.melt():

x = df.melt("Sentiments", var_name="Category")
x = x[x.value != 0].drop(columns="value")
x["Category"] = x["Category"].str.replace(r"s$", "", regex=True)
print(x)

打印：

              Sentiments  Category
0             He is good  Positive
2           She love her  Positive
4   I like going outside  Positive
7              He is bad  Negative
15   She is fine with it   Neutral
17              Its okay   Neutral

【讨论】：

【解决方案2】：

假设只有一列值为 1（即假人），请尝试：

>>> df.set_index("Sentiments").idxmax(axis=1).rename("Category").reset_index()
             Sentiments   Category
0            He is good  Positives
1             He is bad  Negatives
2          She love her  Positives
3   She is fine with it   Neutrals
4  I like going outside  Positives
5              Its okay   Neutrals

【讨论】：

【解决方案3】：

另一种方式：

df = df.set_index('Sentiments').dot(df.columns[1:]).reset_index(name = 'Category')

【讨论】：