根据不建议使用的警告修改OHLC重采样代码 [英] Modify OHLC resample code as per deprecated warning

问题描述

问题:

在处理市场数据并将日内数据重新采样到每日时间范围时，如下所示:

When working with market data and resampling intra-day data to the daily timeframe as follows:

ohlc_dict = {
'Open':'first',
'High':'max',
'Low':'min',
'Last': 'last',
'Volume': 'sum'}

data.resample('1D',how=ohlc_dict).tail().dropna()

                Open    High    Last    Low     Volume
    Timestamp                   
    2016-12-27  163.55  164.18  164.11  163.55  144793.00
    2016-12-28  164.18  164.33  164.22  163.89  215288.00
    2016-12-29  164.44  164.65  164.49  164.27  245538.00
    2016-12-30  164.55  164.56  164.18  164.09  286847.00

似乎给了我我需要的输出(仍然需要验证)...

Which seems to gives me the output I need (still need to verify)...

我收到以下警告:

FutureWarning: how in .resample() is deprecated
the new syntax is .resample(...)..apply(<func>)

问题:

如何使用新语法复制resample代码以与使用apply的当前最佳实践保持一致?

How would this resample code be replicated using the new syntax to align with the current best practice using apply?

我尝试过的事情:

仅以data ['Low']为例:

Just using data['Low'] as an example:

def ohlc (df):
    return df['Low'].min()

data.resample('1D').dropna().apply(ohlc,axis=1).tail(2)

Timestamp
2016-12-29   164.45
2016-12-30   164.26
dtype: float64

没有给我相同的结果，我不确定在哪里插入apply.

Does not give me the same results and Im not sure where to insert the apply.

这是数据的一部分，用于测试是否必填:

Here is a slice of the data to test this with if required:

谢谢

推荐答案

.resample()的作用类似于groupby，因此您可以将该字典传递给resample().agg():

.resample() works like groupby so you can pass that dictionary to resample().agg():

df.resample('1D').agg(ohlc_dict).tail().dropna()
Out: 
              Volume    Last    High    Open     Low
Timestamp                                           
2016-12-27  144793.0  164.11  164.18  163.55  163.55
2016-12-28  215288.0  164.22  164.33  164.18  163.89
2016-12-29  245538.0  164.49  164.65  164.44  164.27
2016-12-30  286847.0  164.18  164.56  164.55  164.09

这篇关于根据不建议使用的警告修改OHLC重采样代码的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！