如何避免Pandas中已分配值被覆盖并打印完整DataFrame

在使用Pandas进行数据块分配时，当前实现能完成各分类的初始分配，但当某块容量已满、将同分类下待处理ID分配到新块时，之前已分配的旧值会被覆盖；另外，当前代码只能打印最后一个D块的分配结果，如何实现打印完整的新DataFrame？

import pandas as pd
data = pd.read_csv("database.csv")

blocks = {"A": 5, "B": 30, "C": 50, "D": 100}
for key, value in blocks.items():
    print("For block " + key + " capacity is " + str(value))
    while True:
        s = input("Which category to be allocated? ")
        dataFrame = data[data['Category'].str.contains(s)]
        if not dataFrame.empty:
            dataFrame['Block'] = key

            # for loop for Sr. Number
            spc = blocks[key]
            i = 1
            for index, row in dataFrame.iterrows():
                if i > spc:
                    break
                dataFrame.loc[index, 'Sr.'] = str(i)
                i += 1
            print(dataFrame)
            break

print(dataFrame)

Name,ID,Category 
ABC,2020,G 
WER,2021,M 
XCV,2022,T 
GFV,2034,M 
WEQ,2021,M 
WEW,2021,M 
WET,2021,M 
WEY,2021,M 
WLK,2021,M 
WYH,2021,M 

Name,ID,Category,Block, Sr. 
1 WER 2021 M A 1 
3 GFV 2034 M A 2 
4 WEQ 2021 M A 3 
5 WEW 2021 M A 4 
6 WET 2021 M A 5 

7 WEY 2021 M B 1
8 WLK 2021 M B 2
9 WYH 2021 M B 3

原代码存在两个核心问题：

• 每次循环都从原始数据中提取整个分类的子集，修改后未同步回原始DataFrame，导致之前的分配结果被覆盖
• 最后仅打印了最后一次循环生成的子集变量，而非完整的处理结果

修改后的代码直接操作原始数据，仅处理未分配的行，保留所有分配记录：

import pandas as pd

# 读取原始数据
data = pd.read_csv("database.csv")
# 初始化Block和Sr.列，避免后续赋值报错
data['Block'] = pd.NA
data['Sr.'] = pd.NA

blocks = {"A": 5, "B": 30, "C": 50, "D": 100}

for block_key, block_capacity in blocks.items():
    print(f"For block {block_key} capacity is {block_capacity}")
    while True:
        category = input("Which category to be allocated? ")
        # 筛选当前分类下未分配Block的行
        unassigned_rows = data[(data['Category'] == category) & data['Block'].isna()]
        if not unassigned_rows.empty:
            # 取最多block_capacity条未分配数据
            assign_count = min(block_capacity, len(unassigned_rows))
            target_indices = unassigned_rows.index[:assign_count]
            
            # 更新原始数据的Block和序号
            data.loc[target_indices, 'Block'] = block_key
            data.loc[target_indices, 'Sr.'] = range(1, assign_count + 1)
            
            # 打印当前块的分配结果
            print(data.loc[target_indices])
            break
        else:
            print(f"该分类{category}已无未分配数据，请重新选择")

# 打印完整的已分配结果
print("\n=== 完整分配结果 ===")
print(data.dropna(subset=['Block']))

关键修改点 #

1. 初始化列：提前创建Block和Sr.列，避免赋值时出现KeyError
2. 筛选未分配行：每次只处理当前分类下还没分配Block的行，彻底解决旧值被覆盖的问题
3. 直接修改原始数据：所有分配操作都在原始data上进行，保留所有历史分配记录
4. 完整结果输出：最后通过dropna筛选出已分配的行，打印完整结果

内容的提问来源于stack exchange，提问作者Adam Burri