熊猫重新采样数据框

Question

I have a resampling (downsampling) problem that should be straightforward to do but I'm not able!! Here is a simplified example:

df:  
       Time         A
0   0.01591  0.108929
1   0.27973  0.411764
2   0.55044  0.064253
3   0.81386  0.317394
4   1.07983  0.722707
5   1.35051  1.154193
6   1.61495  1.151492
7   1.88035  0.123389
8   2.15462  0.093583
9   2.41534  0.260944
10  2.67992  1.007564
11  2.95148  0.325353
12  3.21364  0.555593
13  3.47980  0.740621
15  4.01519  1.619669
16  4.28679  0.477371
17  4.55482  0.432049
18  4.81570  0.194224
19  5.07992  0.331936

The Time column is in seconds. I would like to make the Time column the index and downsample the dataframe to 1s. Help please?

Answer 1

You can use reindex and choose one fill method

In [37]: df.set_index('Time').reindex(range(0,6), method='bfill')
Out[37]: 
          A
0  0.108929
1  0.722707
2  0.093583
3  0.555593
4  1.619669
5  0.331936

Answer 2

First convert your index to datetime format:

df.index=pd.to_datetime(df.Time,unit='s')

Then resample by second (this is the mean value by default but can be changed to sum etc - eg add how='sum' as parameter):

d.resample('S')

                         Time         A
Time                                   
1970-01-01 00:00:00  0.414985  0.225585
1970-01-01 00:00:01  1.481410  0.787945
1970-01-01 00:00:02  2.550340  0.421861
1970-01-01 00:00:03  3.346720  0.648107
1970-01-01 00:00:04  4.418125  0.680828
1970-01-01 00:00:05  5.079920  0.331936

The year/date can be changed if important.