python - 将包含元组列表的列扩展到当前数据帧

Question

我有一个以下格式的数据框:

df = pd.DataFrame({'column_with_tuples': [[('word1', 10), ('word2', 20), ('word3', 30)], [('word4', 40), ('word5', 50), ('word6', 60)]],
                   'category':['category1','category2']})

我想将元组移动到两个单独的列中并保留类别列，以便能够轻松过滤每个类别最常见的单词。

所以最终结果应该是这样的:

df_new = pd.DataFrame({'word': ['word1','word2', 'word3','word4','word5','word6'],
                   'frequency': [10, 20, 30, 40, 50, 60],
                   'category':['category1','category1', 'category1', 'category2', 'category2', 'category2']})

我尝试使用此代码，但结果不是我期望的:

df_tuples = pd.concat([pd.DataFrame(x) for x in df['column_with_tuples']], ignore_index=True)

df_tuples.columns = ['word', 'frequency']

df.drop(['column_with_tuples'], axis=1, inplace=True)

df = pd.concat([df, df_tuples], axis=1)

我希望能得到一些帮助。

Answer 1

您应该使用.explode()方法将 column_with_tuples 列中的元组展开为单独的行。之后，引入 .rename() 方法来更改列的名称，然后将元组解压到单独的列中，并使用 .apply 添加 category 列() 方法。最后用 assign() 方法将 category 列添加到数据帧中。

df_new = df.explode("column_with_tuples")
df_new = df_new.rename(columns={"column_with_tuples": "word"})
df_new[["word", "frequency"]] = df_new["word"].apply(pd.Series)

df_new = df_new.assign(category=df["category"])
df_new = df_new[["word", "frequency", "category"]]
df_new.reset_index(drop=True, inplace=True)
print(df_new)

上述代码的简化版本:

df_new = df.explode("column_with_tuples").rename(columns={"column_with_tuples": "word"})
df_new[["word", "frequency"]] = df_new["word"].apply(pd.Series)
df_new.assign(category=df["category"])

df_new = df_new[["word", "frequency", "category"]].reset_index(drop=True)
print(df_new)

---

    word  frequency   category
0  word1         10  category1
1  word2         20  category1
2  word3         30  category1
3  word4         40  category2
4  word5         50  category2
5  word6         60  category2