Question

我的pandas DataFrame看起来像这样：

             Time         Conc.      Flux
0        0.220000  0.000000e+00       NaN
1        0.536800  0.000000e+00       NaN
2        0.992992  0.000000e+00       NaN
3        1.000000  0.000000e+00 -0.009888
4        1.220000  0.000000e+00       NaN
5        1.536800  0.000000e+00       NaN
6        1.992992  0.000000e+00       NaN
7        2.649909  0.000000e+00       NaN
8        3.595869  0.000000e+00       NaN
9        4.958052  0.000000e+00       NaN
10       6.919595  0.000000e+00       NaN
11       9.744217  0.000000e+00       NaN
12      13.811673  0.000000e+00       NaN
13      19.668812  0.000000e+00       NaN
14      28.103090  0.000000e+00       NaN
15      31.000000  0.000000e+00 -0.009729
16      31.220001  0.000000e+00       NaN
17      31.536800  0.000000e+00       NaN
18      31.992992  0.000000e+00       NaN
19      32.649910  0.000000e+00       NaN
20      33.595871  0.000000e+00       NaN
21      34.958054  0.000000e+00       NaN
22      36.919594  0.000000e+00       NaN
23      39.744217  0.000000e+00       NaN
24      43.811672  0.000000e+00       NaN
25      49.668808  0.000000e+00       NaN
26      58.103088  0.000000e+00       NaN
27      61.000000  0.000000e+00 -0.009751
.               .             .         .
.               .             .         .
.               .             .         .

我试图以有效（快速）的方式将NaN替换为它们下方的数字。基本上，Flux列是一个阶跃函数，它一直保持到下一个值。我确实有数据框与助焊剂分开：

     Time [day]      Flux
0           1.0 -0.009888
1          31.0 -0.009729
2          61.0 -0.009751
3          91.0 -0.009727
4         121.0 -0.009723
5         151.0 -0.016197
6         181.0 -0.015375
7         211.0 -0.014224
8         241.0 -0.019393
9         271.0 -0.012164
.             .         .
.             .         .
.             .         .

我尝试在两个数据帧之间使用嵌套循环并逐个重写NaN，但速度非常慢。我有~100个数据帧，每个数据帧有~4000行。 注意：浓度不一定为零，NaN之间出现数字的距离可能会发生变化。

Answer 1

尝试bfill

df=df.bfill()
df
Out[96]: 
         Time  Conc.      Flux
0    0.220000    0.0 -0.009888
1    0.536800    0.0 -0.009888
2    0.992992    0.0 -0.009888
3    1.000000    0.0 -0.009888
4    1.220000    0.0 -0.009729
5    1.536800    0.0 -0.009729
6    1.992992    0.0 -0.009729
7    2.649909    0.0 -0.009729
8    3.595869    0.0 -0.009729
9    4.958052    0.0 -0.009729
10   6.919595    0.0 -0.009729
11   9.744217    0.0 -0.009729
12  13.811673    0.0 -0.009729
13  19.668812    0.0 -0.009729
14  28.103090    0.0 -0.009729
15  31.000000    0.0 -0.009729
16  31.220001    0.0 -0.009751
17  31.536800    0.0 -0.009751
18  31.992992    0.0 -0.009751
19  32.649910    0.0 -0.009751
20  33.595871    0.0 -0.009751
21  34.958054    0.0 -0.009751
22  36.919594    0.0 -0.009751
23  39.744217    0.0 -0.009751
24  43.811672    0.0 -0.009751
25  49.668808    0.0 -0.009751
26  58.103088    0.0 -0.009751
27  61.000000    0.0 -0.009751

充分填补pandas DataFrame中的空白

1 个答案: