Python 如何比较两列的值并基于比较对值重新排序

我是python新手，我遇到了一个难题，我不知道如何去做。我需要比较两列中的值，然后根据此比较对值重新排序。例如，假设我们有以下数据帧。现在我需要检查B中的值，如果它们小于A的累积和中的值。如果B中的值高于A的cum，那么我们停在那一行。我们根据B列的数值对观测结果重新排序，直到该行（也包括该行）为止

例如，当我们看B中的值时，我们看到所有的值都小于A中的值，直到第4部分

零件号 A. B 一杯 1. 2. 13 2. 2. 4. 17 6. 3. 7. 15 13 4. 5. 16 18 5. 10 19 28 6. 9 16 37 7. 8. 12 45 这应该起作用：

import pandas as pd
data = [ { "Part No": 1, "A": 2, "B": 13, "cum of A": 2 }, { "Part No": 2, "A": 4, "B": 17, "cum of A": 6 }, { "Part No": 3, "A": 7, "B": 15, "cum of A": 13 }, { "Part No": 4, "A": 5, "B": 16, "cum of A": 18 }, { "Part No": 5, "A": 10, "B": 19, "cum of A": 28 }, { "Part No": 6, "A": 9, "B": 16, "cum of A": 37 }, { "Part No": 7, "A": 8, "B": 12, "cum of A": 45 } ]
df = pd.DataFrame(data)
smaller_indexes = len(df[df['B'] > df['cum of A']]) + 1 # check where B is larger than 'cum of A', take the length of that part of the dataframe and add 1 to get the index number of the last value to use for sorting
df[:smaller_indexes] = df[:smaller_indexes].sort_values(by=['B']) # slice the dataframe by the found index number, and sort only that slice of the dataframe
df['cum of A'] = df['A'].cumsum() #recount the cumsum of the entire dataframe

这将输出：

零件号 A. B 一杯 0 1. 2. 13 2. 1. 3. 7. 15 9 2. 4. 5. 16 14 3. 2. 4. 17 18 4. 5. 10 19 28 5. 6. 9 16 37 6. 7. 8. 12 45

---

我知道这有点粗糙，但如果您愿意，您可以使用以下选项：

#order the values in list
def order(a):
    for i in range(len(a)):
        # print(i)
        b = a[i-1] - a[i]
        
        if b>0:
            temp = a[i-1]
            a[i-1] = a[i]
            a[i] = temp
            for j in range(i,0,-1):
                b = a[j-1] - a[j]
                if b>0:
                    temp = a[j-1]
                    a[j-1] = a[j]
                    a[j] = temp
    return a

b = [13,15,16,17,19,16,12]

a = [2, 7, 5, 4, 10, 9, 8]

#other list to append the last value if cum is greater than value of "b" after ordering b
otherList = [] 

# isRecursive and indexToTake is for: 
# "And if there is still a higher value in cum of A than the values in B 
# among these reordered observations (first four parts), then part 4 will be 
# taken out from the list and put in a separate list."

def solve(isRecursive, indexToTake):
    c = 0
    for i in range(len(a)):
        c += a[i] #cum sum
        if c> b[i]: #check if cum is greater than the value of "b" at that index
            if isRecursive:
                if i >indexToTake:
                    return
                otherList.append(b[indexToTake])
                return
            b[0:i+1] = order(b[0:i+1])
            solve(True, i) #check if the values at "b" are greater than cum after ordered until the indexToTake
            break

solve(False, 0) #start solving