如何将 OHLCV 数据重采样为 5 分钟？答案

【问题标题】：How to resample OHLCV data into 5-min?如何将 OHLCV 数据重采样为 5 分钟？
【发布时间】：2018-01-23 23:10:14
【问题描述】：

我有这组数据

2016-08-09 12:39:00,536.7841,536.7849,536.6141,536.7849,0.656
2016-08-09 12:40:00,536.6749,536.6749,536.6749,536.6749,0.2642
2016-08-09 12:41:00,535.84,535.84,535.615,535.615,0.348
2016-08-09 12:42:00,535.5401,535.5401,534.1801,534.1801,0.507
2016-08-09 12:43:00,534.5891,534.8753,534.5891,534.807,0.656
2016-08-09 12:44:00,534.8014,534.878,534.8014,534.8416,0.502
2016-08-09 12:45:00,534.8131,534.8131,534.2303,534.6736,0.552
2016-08-09 12:47:00,534.756,538.5999,534.756,534.7836,0.62647241
2016-08-09 12:48:00,536.0557,536.6864,536.0557,536.6864,1.2614
2016-08-09 12:49:00,536.8966,537.7289,536.8966,537.7289,0.532
2016-08-09 12:50:00,537.9829,539.2199,537.9829,539.2199,0.67752932
2016-08-09 12:51:00,538.5,539.2199,538.5,539.2199,0.43768953

我想将其重新采样为 5 分钟 OHCLV，所以我做了以下代码：

import pandas as pd

df= pd.read_csv("C:\Users\Araujo's PC\Desktop\python_scripts\CSV\cex_btc.csv",
                names=['timestamps','open','high','low','close','volume'])

df.set_index('timestamps',inplace=True)

ohlc_dict = {
    'open':'first',
    'high':'max',
    'low':'min',
    'close':'last',
    'volume':'sum'
    }

df.resample('5T', how=ohlc_dict)

print df

我觉得这个错误：

TypeError: 仅对 DatetimeIndex、TimedeltaIndex 或 PeriodIndex 有效，但获得了 'Index' 的实例

有人可以帮我解决这个问题吗？

【问题讨论】：

Resampling Minute data的可能重复

标签： python dataframe data-analysis sampling trading

【解决方案1】：

您只需将timestamps 列中的值转换为pandas 时间戳，然后再使用它们的值设置索引。我相信它们目前只是文本字段。

df['timestamps'] = pd.to_datetime(df['timestamps'])
df.set_index('timestamps', inplace=True)

>>> df.resample('5T', how=ohlc_dict)

                         high     close      open       low    volume
timestamps                                                           
2016-08-09 12:35:00  536.7849  536.7849  536.7841  536.6141  0.656000
2016-08-09 12:40:00  536.6749  534.8416  536.6749  534.1801  2.277200
2016-08-09 12:45:00  538.5999  537.7289  534.8131  534.2303  2.971872
2016-08-09 12:50:00  539.2199  539.2199  537.9829  537.9829  1.115219

您也可以在读取 csv 时尝试解析这些：

pd.read_csv(filename, parse_dates=['timestamps'],
            names=['timestamps','open','high','low','close','volume'])

【讨论】：