pandas - 将 pandas Dataframe 转换为多个 numpy 数组

Question

我想将我的列转换为多个数组，这样每一行都有来自同一日期的值

My data:

Date        Value
2006-12-11  816.2
2006-12-11  816.2
2006-12-11  816.2
2006-12-12  848.2
2006-12-12  849.0
2006-12-13  885.6
2006-12-13  887.0
2006-12-13  885.2
2006-12-13  882.0
2006-12-13  885.0

预期输出:

[[816.2, 816.2, 816.2]
[848.2, 849.0]
[885.6, 887.0, 885.2, 882.0, 885.0]] 

Answer 1

您可以将 groupby 与 numpy.array 一起使用:

import numpy as np:

df.Value.groupby(df.Date).apply(np.array).values

例子:

df = pd.DataFrame({
    "Date": ['2006-12-11', '2006-12-11', '2006-12-11', '2006-12-12', '2006-12-12'], 
    'Value': [816.2, 816.2, 816.2, 848.2, 849.0]})
df.Value.groupby(df.Date).apply(np.array).values
array([array([816.2, 816.2, 816.2]), array([848.2, 849. ])], dtype=object)