Python Pandas 如何比较一个 Dataframe 中的日期与另一个 ZC699575A5E8AFD9E22A7ECC8CAB 中的日期？

Question

我有 Dataframe 1：

Hotel   DateFrom    DateTo      Room
BBB     2019-10-29  2020-03-27  DHS
BBB     2020-03-28  2020-10-30  DHS
BBB     2020-10-31  2021-03-29  DHS
BBB     2021-03-30  2099-01-01  DHS

和 Dataframe 2：

Hotel   DateFrom    DateTo      Room    Food
BBB     2020-03-01  2020-04-24  DHS     A
BBB     2020-04-25  2020-05-03  DHS     B
BBB     2020-05-04  2020-05-31  DHS     C
BBB     2020-06-01  2020-06-22  DHS     D
BBB     2020-06-23  2020-08-26  DHS     E
BBB     2020-08-27  2020-11-30  DHS     F

我需要检查 df1 中的每一行是否以及 df1_DateFrom 是否介于 df2_DateFrom 和 df2_DateTo 之间。 然后我需要将该食品代码从 df2 获取到 df1 中的新列或如下所示的新 df3。

结果将如下所示：

df3:

    Hotel   DateFrom    DateTo      Room  Food
    BBB     2019-10-29  2020-03-27  DHS   
    BBB     2020-03-28  2020-10-30  DHS   A
    BBB     2020-10-31  2021-03-29  DHS   F 
    BBB     2021-03-30  2099-01-01  DHS

我真的很感激这方面的任何帮助。 我在 Pandas 上有点新，还在学习，我必须说这对我来说有点复杂。

Answer 1

您可以进行交叉合并和查询：

# recommend dealing with datetime type:
df1['DateFrom'],df1['DateTo'] = pd.to_datetime(df1['DateFrom']),pd.to_datetime(df1['DateTo'])
df2['DateFrom'],df2['DateTo'] = pd.to_datetime(df2['DateFrom']),pd.to_datetime(df2['DateTo'])

new_df = (df1.reset_index().merge(df2, on=['Hotel','Room'],
                                  how='left', suffixes=['','_'])
             .query('DateFrom_ <= DateFrom <= DateTo_')
         )
df1['Food'] = new_df.set_index('index')['Food']

Output：

  Hotel   DateFrom     DateTo Room Food
0   BBB 2019-10-29 2020-03-27  DHS  NaN
1   BBB 2020-03-28 2020-10-30  DHS    A
2   BBB 2020-10-31 2021-03-29  DHS    F
3   BBB 2021-03-30 2099-01-01  DHS  NaN

Answer 2

远不如 Quang Hoang 的回答优雅，但使用np.piecewise的解决方案看起来像这样。 另请参阅https://stackoverflow.com/a/30630905/4873972

import pandas as pd
import numpy as np
from io import StringIO

# Creating the dataframes.
df1 = pd.read_table(StringIO("""
Hotel   DateFrom    DateTo      Room
BBB     2019-10-29  2020-03-27  DHS
BBB     2020-03-28  2020-10-30  DHS
BBB     2020-10-31  2021-03-29  DHS
BBB     2021-03-30  2099-01-01  DHS
"""), sep=r"\s+").convert_dtypes()

df1["DateFrom"] = pd.to_datetime(df1["DateFrom"])
df1["DateTo"] = pd.to_datetime(df1["DateTo"])

df2 = pd.read_table(StringIO("""
Hotel   DateFrom    DateTo      Room    Food
BBB     2020-03-01  2020-04-24  DHS     A
BBB     2020-04-25  2020-05-03  DHS     B
BBB     2020-05-04  2020-05-31  DHS     C
BBB     2020-06-01  2020-06-22  DHS     D
BBB     2020-06-23  2020-08-26  DHS     E
BBB     2020-08-27  2020-11-30  DHS     F
"""), sep=r"\s+").convert_dtypes()

df2["DateFrom"] = pd.to_datetime(df2["DateFrom"])
df2["DateTo"] = pd.to_datetime(df2["DateTo"])
# Avoid zero index for merging later on.
df2["id"] = np.arange(1, len(df2) +1 )

# Find matching indexes.
df1["df2_id"] = np.piecewise(
    np.zeros(len(df1)), 
    [(df1["DateFrom"].values >= start_date) & (df1["DateFrom"].values <= end_date) for start_date, end_date in zip(df2["DateFrom"].values, df2["DateTo"].values)], 
    df2.index.values
)

# Merge on matching indexes.
df1.merge(df2["Food"], left_on="df2_id", right_index=True, how="left")

Output：

  Hotel   DateFrom     DateTo Room Food
0   BBB 2019-10-29 2020-03-27  DHS  NaN
1   BBB 2020-03-28 2020-10-30  DHS    A
2   BBB 2020-10-31 2021-03-29  DHS    F
3   BBB 2021-03-30 2099-01-01  DHS  NaN