如何根据另一个数据框的条件从数据框中选择行答案

【问题标题】：How to select rows from a dataframe based on conditions with another dataframe如何根据另一个数据框的条件从数据框中选择行
【发布时间】：2023-01-07 00:53:09
【问题描述】：

我有两个这样的数据框

df1:
Date       Fruit   
2022-11-24 Banana 
2021-5-23  Orange  
2020-10-1  Apple

df2:
Start_Date End_Date    Fruit
2021-11-20 2022-12-31  Banana 
2022-6-1   2022-10-31  Banana   
2013-14-1  2022-11-31  Banana 
2013-11-24 2022-10-31  Celery 
2013-11-25 2022-10-31  Apple  
2013-11-25 2022-10-31  Orange

我想创建一个数据框 df3，它由 df2 中的行组成，条件如下：

df1中有一行Fruit值相同，df1的值Date在df2的Start_Date和End_Date之间

所以对于这个例子，df3 将是：

df3:
Start_Date End_Date    Fruit
2021-11-20 2022-12-31  Banana    
2013-14-01 2022-11-31  Banana  
2013-11-25 2022-10-31  Apple  
2013-11-25 2022-10-31  Orange

我试过了，但它似乎不是更优化的方式，并且需要时间来执行大型数据帧

df3 = pd.DataFrame()

for _, row in df1.iterrows():
    FRUIT= row['FRUIT']
    DATE= row['DATE']
    temp= df2[(df2['FRUIT'] == FRUIT) &(df2['Start_Date'] <= DATE) &(df2['End_Date'] >= DATE)] 
    
    df3 = pd.concat([df3 , temp])

谢谢您的帮助！

【问题讨论】：

标签： python dataframe for-loop

【解决方案1】：

import pandas as pd

# create a list to store the rows of df3
df3_rows = []

# iterate through the rows of df1
for _, row in df1.iterrows():
  date = row['Date']
  fruit = row['Fruit']
  
  # use boolean indexing to select rows from df2 that meet the conditions
  df2_subset = df2[(df2['Fruit'] == fruit) & (df2['Start_Date'] <= date) & (df2['End_Date'] >= date)]
  
  # add the rows from df2_subset to df3_rows
  df3_rows.extend(df2_subset.to_dict('records'))

# create df3 from the list of rows
df3 = pd.DataFrame(df3_rows)

【讨论】：

【解决方案2】：

#For 2013-14-1, there is no month 14, it would get error. 
#And for 2022-11-31, there is no day 31 for Nov. 
#Please correct these two records before running the following code.

df1=pd.read_csv("test1.txt",parse_dates=[0],sep='	')
df2=pd.read_csv("test2.txt",parse_dates=[0,1],sep='	') 
#or you can also use pd.to_datetime to convert string into DateTime.
print(df1.dtypes)
print(df2.dtypes)
   
df3=df2.assign(Date=df2.Fruit.map(df1.set_index('Fruit').Date.to_dict())).dropna(subset='Date')
df3=df3.loc[df3.apply(lambda x:x.Date >= x.Start_Date and x.Date <= x.End_Date,axis=1)]
print(df3)

输出：

Date     datetime64[ns]
Fruit    object
dtype: object

Start_Date    datetime64[ns]
End_Date      datetime64[ns]
Fruit         object        
dtype: object


  Start_Date   End_Date   Fruit       Date
0 2021-11-20 2022-12-31  Banana 2022-11-24
2 2013-12-01 2022-11-30  Banana 2022-11-24
4 2013-11-25 2022-10-31  Apple  2020-10-01
5 2013-11-25 2022-10-31  Orange 2021-05-23

【讨论】：