非固定滚动窗口

本文介绍了非固定滚动窗口的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！ 问题描述

我希望在列表上实现滚动窗口，但我想提供一个滚动窗口列表，而不是固定长度的窗口:像这样:

I am looking to implement a rolling window on a list, but instead of a fixed length of window, I would like to provide a rolling window list: Something like this:

l1 = [5, 3, 8, 2, 10, 12, 13, 15, 22, 28] l2 = [1, 2, 2, 2, 3, 4, 2, 3, 5, 3] get_custom_roling( l1, l2, np.average)

结果是:

[5, 4, 5.5, 5, 6.67, ....]

6.67 计算为 3 个元素 10、2、8 的平均值.

6.67 is calculated as average of 3 elements 10, 2, 8.

我实施了一个缓慢的解决方案，欢迎任何想法使其更快:):

I implemented a slow solution, and every idea is welcome to make it quicker :):

import numpy as np def get_the_list(end_point, number_points): """ example: get_the_list(6, 3) ==> [4, 5, 6] example: get_the_list(9, 5) ==> [5, 6, 7, 8, 9] """ if np.isnan(number_points): return [] number_points = int( number_points) return list(range(end_point, end_point - number_points, -1 )) def get_idx(s): ss = list(enumerate(s) ) sss = (get_the_list(*elem) for elem in ss ) return sss def get_custom_roling(s, ss, funct): output_get_idx = get_idx(ss) agg_stuff = [s[elem] for elem in output_get_idx] res_agg_stuff = [ funct(elem) for elem in agg_stuff ] res_agg_stuff = eiu.pd.Series(data=res_agg_stuff, index = s.index) return res_agg_stuff

推荐答案

Pandas 自定义窗口滚动 允许您修改窗口大小.

Pandas custom window rolling allows you to modify size of window.

简单解释:start 和 end 数组保存索引值以制作数据切片.

Simple explanation: start and end arrays hold values of indexes to make slices of your data.

#start = [0 0 1 2 2 2 5 5 4 7] #end = [1 2 3 4 5 6 7 8 9 10]

传递给 get_window_bounds 的参数由 BaseIndexer 给出.

Arguments passed to get_window_bounds are given by BaseIndexer.

import pandas as pd import numpy as np from pandas.api.indexers import BaseIndexer from typing import Optional, Tuple class CustomIndexer(BaseIndexer): def get_window_bounds(self, num_values: int = 0, min_periods: Optional[int] = None, center: Optional[bool] = None, closed: Optional[str] = None ) -> Tuple[np.ndarray, np.ndarray]: end = np.arange(1, num_values+1, dtype=np.int64) start = end - np.array(self.custom_name_whatever, dtype=np.int64) return start, end

df = pd.DataFrame({"l1": [5, 3, 8, 2, 10, 12, 13, 15, 22, 28], "l2": [1, 2, 2, 2, 3, 4, 2, 3, 5, 3]}) indexer = CustomIndexer(custom_name_whatever=df.l2) df["variable_mean"] = df.l1.rolling(indexer).mean() print(df)

输出:

l1 l2 variable_mean 0 5 1 5.000000 1 3 2 4.000000 2 8 2 5.500000 3 2 2 5.000000 4 10 3 6.666667 5 12 4 8.000000 6 13 2 12.500000 7 15 3 13.333333 8 22 5 14.400000 9 28 3 21.666667