能用 nosql 吗

如果 A/B 没规律的话，可以用自定义 window rolling
from pandas.api.indexers import BaseIndexer
window_size = df.C.groupby((df.C != df.C.shift(1)).cumsum()).agg('sum').max() # 最大连续次数
class CustomIndexer(BaseIndexer):
def get_window_bounds(self, num_values, min_periods, center, closed):
start = np.empty(num_values, dtype=np.int64)
end = np.empty(num_values, dtype=np.int64)
for i in range(num_values):
end[i] = i + 1
j = i
while j > 0 and self.use_expanding[j]:
j -= 1
start[i] = j
return start, end
indexer = CustomIndexer(window_size=window_size, use_expanding=df.C)
res = df.B.rolling(indexer, min_periods=2).sum().fillna(0)

标准库 functools.lru_cache 把它当双向链表用
摘部分代码：
PREV, NEXT, KEY, RESULT = 0, 1, 2, 3 # names for the link fields
root = [] # root of the circular doubly linked list
root[:] = [root, root, None, None] # initialize by pointing to self

# Use the old root to store the new key and result.
oldroot = root
oldroot[KEY] = key
oldroot[RESULT] = result
# Empty the oldest link and make it the new root.
# Keep a reference to the old key and old result to
# prevent their ref counts from going to zero during the
# update. That will prevent potentially arbitrary object
# clean-up code (i.e. __del__) from running while we're
# still adjusting the links.
root = oldroot[NEXT]
oldkey = root[KEY]
oldresult = root[RESULT]
root[KEY] = root[RESULT] = None

各位想复杂了，pandas.read_csv 有现成的 chunksize 参数

国产车比国产程序员质量好一点

pd.read_excel 有个可选参数 parse_cols

greenlet.getcurrent 。
你的需求和 flask 的实现好像，可以看看 werkzeug.local 的源码

先体验无休~

你可以用 collections.ChainMap，或者参考 flask 里的 MultiDict：
>>> d = MultiDict([('a', 'b'), ('a', 'c')])
>>> d
MultiDict([('a', 'b'), ('a', 'c')])
>>> d['a']
'b'
>>> d.getlist('a')
['b', 'c']

df['diff'] = df.high.shift() - df.high
df[df.diff > 0][df.markf == -1]

Marshmallow/Pydantic/JSON Schema

财大气粗，德高望重

你需要创建缓存实例
python cookbook 里的例子：

import weakref

class Cached(type):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.__cache = weakref.WeakValueDictionary()

def __call__(self, *args):
if args in self.__cache:
return self.__cache[args]
else:
obj = super().__call__(*args)
self.__cache[args] = obj
return obj

# Example
class Spam(metaclass=Cached):
def __init__(self, name):
self.name = name


>>> a = Spam('Guido')
>>> b = Spam('Diana')
>>> c = Spam('Guido') # Cached
>>> a is b
False
>>> a is c # Cached value returned
True

用队列。。。

1.最新的 pyarrow 是支持 nested data 的： https://github.com/apache/arrow/pull/6751/files
2.这时 pandas dataframe 转 csv 没问题：
s = StringIO()
df = pd.DataFrame({'c1': [[1, 2]], 'c2': [{'k': 'v'}]})
df.to_csv(s, index=False)
s.seek(0)
s.read() # 'c1,c2\r\n"[1, 2]",{\'k\': \'v\'}\r\n'

楼主同事买房和他积蓄 15 万一点关系都没有，和他月薪 3w 公积金 6k 有关系。
标题不写成“月薪 3w 公积金 6k 积蓄在郊区买房的故事，买房真没那么难”，而是“15w 积蓄在郊区买房的故事，买房真没那么难”，这是偷换概念，挨骂一点不冤。

可以参考 werkzeug.datastructures 里实现的 ImmutableDict
flask 就用到了
class Flask(_PackageBoundObject):
...
default_config = ImmutableDict({"ENV": None, ...})

def stats(callback_name, **tags):
for instance_id in get_category_instance_id_by_tags(**tags):
ecs_obj = ECSInstance(instance_id)
yield getattr(ecs_obj, callback_name)


def cal_category_ecs_cpu_core_amount_by_tags(**tags):
return sum(stats('get_cpu_core', **tags))


def cal_category_ecs_memory_amount_by_tags(**tags):
return sum(stats('get_memory', **tags))


其实应该修改的是 ECSInstance