Python pandas interpolating series

问题

I have data in a csv file which appears as:

  DateTime       Temp
10/1/2016 0:00  20.35491156
10/1/2016 1:00  19.75320845
10/1/2016 4:00  17.62411292
10/1/2016 5:00  18.30190001
10/1/2016 6:00  19.37101638

I am reading this file from csv file as:

import numpy as np
import pandas as pd
d2 = pd.Series.from_csv(r'C:\PowerCurve.csv')
d3 = d2.interpolate(method='time')

My goal is to fill the missing hours 2 and 3 with interpolation based on nearby values. i.e. every time there is are missing data it should do the interpolation.

However, d3 doesn't show any interpolation.

Edit: Based on suggestions below my Python 2.7 still errors out. I am trying the following:

import pandas as pd
d2 = pd.Series.from_csv(r'C:\PowerCurve.csv')
d2.set_index('DateTime').resample('H').interpolate()

Error is:

File "C:\Python27\lib\site-packages\pandas\core\generic.py", line 2672, in __getattr__
    return object.__getattribute__(self, name)
AttributeError: 'Series' object has no attribute 'set_index'

回答1:

use the interpolate method after resample on an hourly basis.

d2.set_index('DateTime').resample('H').interpolate()

If d2 is a series then we don't need the set_index

d2.resample('H').interpolate()

回答2:

Use resample with datetime as index and use one of the methods of resampling that fits your need. For instance:

df.set_index('DateTime').resample('1H').pad()
Out[23]: 
                          Temp
DateTime                      
2016-10-01 00:00:00  20.354912
2016-10-01 01:00:00  19.753208
2016-10-01 02:00:00  19.753208
2016-10-01 03:00:00  19.753208
2016-10-01 04:00:00  17.624113
2016-10-01 05:00:00  18.301900
2016-10-01 06:00:00  19.371016

来源：https://stackoverflow.com/questions/39949897/python-pandas-interpolating-series

标签

python

pandas

interpolation