如何合并含字符串、整数与元组的混合列表元素:按所属方聚合统计
混合列表按所属方聚合Gizmo数值的实现
原代码报错的根因
- 直接将装着Owner字符串的单元素列表(如
['Owner1'])作为字典键:列表是可变、不可哈希的类型,本身不满足字典键的基本要求 - 累加逻辑没有提取实际数值:循环中拿到的
val是嵌套结构[('Gizmo1', 1.0)],既没有取出内层的元组,也没有拆分出元组里的浮点数量值,直接将整个嵌套结构和默认初始值0做加法,最终触发int与tuple类型无法运算的报错
可直接运行的实现方案
无需拆分原列表的简化方案
不需要提前把原混合列表拆成items_qty和owner两个子列表,直接遍历原结构即可完成聚合:
results = [[('Gizmo1', 1.0), 'Owner1'], [('Gizmo2', 4.0), 'Owner2'],[('Gizmo1', 4.0), 'Owner1'],[('Gizmo2', 4.0), 'Owner2']] aggregate = {} for item in results: gizmo_meta, owner_name = item gizmo_name, qty = gizmo_meta # 用(Gizmo名称, 所属方)作为聚合键,*避免同个所属方下不同Gizmo的数值被错误合并* agg_key = (gizmo_name, owner_name) aggregate[agg_key] = aggregate.get(agg_key, 0) + qty # 转换为要求的输出结构 final_results = [[(gizmo_name, total_qty), owner_name] for (gizmo_name, owner_name), total_qty in aggregate.items()]
运行后输出结果与预期完全一致:
[[('Gizmo1', 5.0), 'Owner1'], [('Gizmo2', 8.0), 'Owner2']]
基于defaultdict的精简写法
如果允许使用标准库组件,可通过collections.defaultdict省去手动判断默认值的步骤,代码更简洁:
from collections import defaultdict results = [[('Gizmo1', 1.0), 'Owner1'], [('Gizmo2', 4.0), 'Owner2'],[('Gizmo1', 4.0), 'Owner1'],[('Gizmo2', 4.0), 'Owner2']] aggregate = defaultdict(float) for item in results: (gizmo_name, qty), owner_name = item aggregate[(gizmo_name, owner_name)] += qty final_results = [[(gizmo_name, total), owner] for (gizmo_name, owner), total in aggregate.items()]
适配原有拆分思路的修改方案
如果需要沿用之前拆分两个子列表的逻辑,需要修正两处核心问题:
- 从单元素嵌套列表中取出实际的字符串型Owner值,不要直接将列表作为字典键
- 从嵌套结构中拆出Gizmo元组、提取对应数值后再执行累加
修改后的可运行代码:
items_qty = [[('Gizmo1', 1.0)], [('Gizmo2', 4.0)],[('Gizmo1', 4.0)],[('Gizmo2', 4.0)]] owner = [['Owner1'], ['Owner2'],['Owner1'],['Owner2']] data = {} for key_wrap, val_wrap in zip(owner, items_qty): owner_key = key_wrap[0] gizmo_name, qty = val_wrap[0] agg_key = (gizmo_name, owner_key) data[agg_key] = data.get(agg_key, 0) + qty final_results = [[(gizmo_name, total_qty), owner] for (gizmo_name, owner), total_qty in data.items()]
内容的提问来源于stack exchange,提问作者apexprogramming
相关产品推荐
相关产品推荐

