python - 我收到零和​​负日 python 和 pandas csv 文件

Question

是否可以使用pandas库替换csv文件中的那些零和负数天，我只需要用零或三十天替换任何负数天

这是我的示例:

import datetime
from dateutil.relativedelta import relativedelta
import pandas as pd
import numpy as np

df = pd.read_csv("csv_data.csv", encoding="UTF-8",
    parse_dates = [
        "firstdate",
        "lastdate",
    ], dayfirst=True
)
diff = pd.firstdate - pd.lastdate
#When i skip the following loop
"""for d in diff:
    if d < 0:
        df['thedif'] = 0
    else:
        df['thedif'] = diff"""
#and add this line below

  df['thedif'] = diff

index firstdate   lastdate   thedif
0     2021-03-02  2021-04-02 -31days
1     2021-04-02  2021-03-02 31days
2     2021-03-03  2021-03-03 0days

但是如果我们在 numpy 中使用 where 关键字，我们会得到

Invalid comparison between dtype=timedelta64[ns] and int

Answer 1

Tou 可以通过 Series.dt.days 将天数转换为整数:

diff = (df.firstdate - df.lastdate).dt.days

df['thedif'] = np.where(diff < 0, 0, diff)

或者通过pd.Timedelta(0)进行比较:

diff = df.firstdate - df.lastdate

df['thedif'] = np.where(diff < pd.Timedelta(0), pd.Timedelta(0), diff)