Pandas 获取数据框的一部分并对值进行归一化

Question

如下两列的数据框。

我想通过给出日期来选择部分，并标准化（通过使用最小-最大方法）"Weight"。

这是我的计划：

import pandas as pd


data = {'Date': ["2000-02-01", "2000-03-01", "2000-04-03", "2000-05-01", "2000-06-01", "2000-07-03", "2000-08-01", "2000-09-01", "2000-10-02", "2000-11-01"], 
'Weight' : [478, 26, 144, 9, 453, 24, 383, 314, 291, 286]}

df = pd.DataFrame(data)

df_1 = df.loc[df['Date'] >= "2000-04-01"]

df_1 = (df_1 - df_1.min()) / (df_1.max() - df_1.min())

print df_1

# the ideal output is two columns: 1 for Dates after "2000-04-01". 1 for their correspondent normalized "Weights".

给出错误：

TypeError: unsupported operand type(s) for -: 'str' and 'str'

如何实现？谢谢。

Answer 1

首先将值转换为日期时间，然后仅处理 Weight 列并覆盖 Weight 列：

df['Date']  = pd.to_datetime(df['Date'] )
df_1 = df.loc[df['Date'] >= "2000-04-01"]

a = (df_1['Weight'] - df_1['Weight'].min()) / (df_1['Weight'].max() - df_1['Weight'].min())

print (df_1.assign(Weight = a))
        Date    Weight
2 2000-04-03  0.304054
3 2000-05-01  0.000000
4 2000-06-01  1.000000
5 2000-07-03  0.033784
6 2000-08-01  0.842342
7 2000-09-01  0.686937
8 2000-10-02  0.635135
9 2000-11-01  0.623874

Answer 2

日期列的数据类型是string.so你必须改变它to.for你可以使用这种方法==>

df['Date']=pd.to_datetime(df['Date'])

Pandas 获取数据框的一部分并对值进行归一化

Pandas to take portions of data-frame and normalize values

python

normalization

dataframe

pandas