透视重复的时间序列数据

Question

我正在尝试以这种方式来透视数据，以使我得到诸如AK_positive AK_probableCases，AK_negative，AL_positive ..等列。

您可以在此处获取数据，df = pd.read_csv（'https://covidtracking.com/api/states/daily.csv'）

Answer 1

只需使用.to_flat_index()将原始MultiIndex列展平为元组，然后将元组元素重新排列为新的列名。

df_pivoted.columns = [f"{i[1]}_{i[0]}" for i in df_pivoted.columns.to_flat_index()]

结果：

# start from April
df_pivoted[df_pivoted.index >= 20200401].head(5)

          AK_positive  AL_positive  AR_positive  ...  WI_grade  WV_grade  WY_grade
date                                             ...                              
20200401        133.0       1077.0        584.0  ...       NaN       NaN       NaN
20200402        143.0       1233.0        643.0  ...       NaN       NaN       NaN
20200403        157.0       1432.0        704.0  ...       NaN       NaN       NaN
20200404        171.0       1580.0        743.0  ...       NaN       NaN       NaN
20200405        185.0       1796.0        830.0  ...       NaN       NaN       NaN