根据 pandas 中另一个数据帧中的某些条件将值从一个数据帧拆分到另一个数据帧

Question

I have two data frames df1 and df2, and I want to adjust values from df2 to df1 based on conditions in df1. The conditions are based on 4 different columns and different conditions for different IDs in df1 and I need to put values of one column from df2 in such a way that it splits the value from df2 and adjust it in df1, and the sum of each IDs value should match in both the data frames.

So I have the following format of data:

I want to bring values from df2 to df1 and split it according to Start Day End Day, Start Time and End Time in df1 itself and the sum should be equal in both df1 and df2 for each ID.

Expected Output

here is the same data frame created in pandas. these two tables are input values and I want the expected result as above.

df1 = pd.DataFrame({'ID': ["Ch1","Ch1","Ch1","Ch1","Ch1","Ch1","Ch2","Ch2","Ch2"],
               'Start Day': [1,1,1,6,6,6,1,1,1], 
               'End Day': [5,5,5,7,7,7,7,7,7], 
               'Start Time': [600,1200,1700,600,1200,1700,700,1200,1700], 
               'End Time': [1200,1700,2500,1200,1700,2500,1200,1700,2400]})
print(df1)

df2 = pd.DataFrame({'ID': ["Ch1","Ch1","Ch1","Ch2","Ch2","Ch2","Ch2","Ch2","Ch2","Ch2","Ch2","Ch2","Ch2","Ch2"],
                    'Start Day': [1,1,1,1,1,1,1,1,1,1,1,1,6,1],
                    'End Day': [7,7,7,5,5,5,5,5,5,5,5,5,7,7],
                    'Start Time': [600,1200,1700,800,900,1000,1100,1200,1300,1900,2000,2200,700,700],
                    'End Time': [1200,1700,2500,900,1000,1100,1200,1300,1400,2000,2200,2300,2400,2400],
                    'Values':[1125,2250,1125,346.5,346.5,346.5,346.5,346.5,346.5,189,189,346.5,1795.5,346.5]})
print(df2)

Can somebody please help me with this.

Thanks in advance.

Answer 1

Like @Tomer S suggested, we really need some more input into how you need the "Values" to be adjusted from DF2 to DF1.

Here is an approach to take the sum of the values from DF2 that match on CH2 and fall within the same window of time defined by the Start/End Days/Times:

def getValueFromDF2(df1row):
    return df2[
        (df2["ID"] == df1row["ID"]) 
        & (df2["Start Day"] >= df1row["Start Day"]) & (df2["End Day"] <= df1row["End Day"])
        & (df2["Start Time"] >= df1row["Start Time"]) & (df2["End Time"] <= df1row["End Time"])
    ]["Values"].sum()

df1["Values"] = df1.apply(lambda row: getValueFromDF2(row), axis=1)
print(df1)

Output:

    ID  Start Day  End Day  Start Time  End Time  Values
0  Ch1          1        5         600      1200     0.0
1  Ch1          1        5        1200      1700     0.0
2  Ch1          1        5        1700      2500     0.0
3  Ch1          6        7         600      1200     0.0
4  Ch1          6        7        1200      1700     0.0
5  Ch1          6        7        1700      2500     0.0
6  Ch2          1        7         700      1200  1386.0
7  Ch2          1        7        1200      1700   693.0
8  Ch2          1        7        1700      2400   724.5

I don't think this is what you want though.
I'm not sure how the sum of values in your output can equal the sum of values in df2, as you stated should be the case, regardless of how we aggregate or split them up, given that the Channel, Day, and Time windows in df2 are not all represented in df1 and vice-versa.