我正在尝试用 pandas 填充 ascii 文件中时间序列数据中缺失的点。我认为其他事情都还好，但是即使最初有数据，第一行也充满了 nan 。 我的数据样本是:
"2011-08-26 00:00:00",1155179,3.232,23.7,3.281,0.386,25.27,111.5665,28.92,29.83,19.13,0,111.5,13.02,29.77,345.7
"2011-08-26 00:00:30",1155180,3.289,20.44,2.153,0.222,25.25,111.5735,28.94,29.82,19.53,0,111.5,13.02,29.79,342.4
.
.
"2011-08-26 23:59:30",1155297,12.62,28.06,3.162,1.356,24.3,111.4614,28.65,29.84,19.53,0,111.4,13.06,29.50,350.1
我使用的代码如下:
t1 = np.genfromtxt(INPUT,dtype=None,delimiter=',',usecols=[0])
start = t1[0].strip('\'"')
end = t1[-1].strip('\'"')
data=pd.read_csv(INPUT,sep=',',index_col=[0],parse_dates=[0])
index = pd.date_range(start,end,freq="30S")
df = data
sk_f = df.reindex(index)
因此，通过这段代码，我想读取第一列的第一个和最后一个字符串，并将它们放入索引以填充表示为 nan 的可能缺失点。但是，问题是第一列也填写了结果，如下所示:
2011-08-26 00:00:00,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan
2011-08-26 00:00:30,1155180,3.289,20.44,2.153,0.222,25.25,111.5735,28.94,29.82,19.53,0,111.5,13.02,29.79,342.4
.
.
2011-08-26 23:59:30,1155297,12.62,28.06,3.162,1.356,24.3,111.4614,28.65,29.84,19.53,0,111.4,13.06,29.50,350.1
这意味着尽管原始文件中有数据，但第一行被意外填充。从第二行开始，一切都很好，填充缺失的数据似乎也很好。我正在尝试找出为什么会发生这种情况。说实话，我现在还没有找到原因。 任何想法或帮助将非常感激。 谢谢你， 艾萨克
最佳答案
我认为您可以省略 genfromtxt 读取文件并仅尝试 read_csv ，然后找到min和 max日期 reindex方法。
或者使用resample :
import pandas as pd
import numpy as np
import io
temp=u""""2011-08-26 00:00:00",1155179,3.232,23.7,3.281,0.386,25.27,111.5665,28.92,29.83,19.13,0,111.5,13.02,29.77,345.7
"2011-08-26 00:00:30",1155180,3.289,20.44,2.153,0.222,25.25,111.5735,28.94,29.82,19.53,0,111.5,13.02,29.79,342.4
"2011-08-26 23:59:30",1155297,12.62,28.06,3.162,1.356,24.3,111.4614,28.65,29.84,19.53,0,111.4,13.06,29.50,350.1"""
#after testing replace io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), sep=",", index_col=[0], parse_dates=[0], header=None)
print df
1 2 3 4 5 6 7 \
0
2011-08-26 00:00:00 1155179 3.232 23.70 3.281 0.386 25.27 111.5665
2011-08-26 00:00:30 1155180 3.289 20.44 2.153 0.222 25.25 111.5735
2011-08-26 23:59:30 1155297 12.620 28.06 3.162 1.356 24.30 111.4614
8 9 10 11 12 13 14 15
0
2011-08-26 00:00:00 28.92 29.83 19.13 0 111.5 13.02 29.77 345.7
2011-08-26 00:00:30 28.94 29.82 19.53 0 111.5 13.02 29.79 342.4
2011-08-26 23:59:30 28.65 29.84 19.53 0 111.4 13.06 29.50 350.1
start = df.index.min()
end = df.index.max()
print start
2011-08-26 00:00:00
print end
2011-08-26 23:59:30
index = pd.date_range(start,end,freq="30S")
sk_f = df.reindex(index)
print sk_f.head()
1 2 3 4 5 6 7 \
2011-08-26 00:00:00 1155179 3.232 23.70 3.281 0.386 25.27 111.5665
2011-08-26 00:00:30 1155180 3.289 20.44 2.153 0.222 25.25 111.5735
2011-08-26 00:01:00 NaN NaN NaN NaN NaN NaN NaN
2011-08-26 00:01:30 NaN NaN NaN NaN NaN NaN NaN
2011-08-26 00:02:00 NaN NaN NaN NaN NaN NaN NaN
8 9 10 11 12 13 14 15
2011-08-26 00:00:00 28.92 29.83 19.13 0 111.5 13.02 29.77 345.7
2011-08-26 00:00:30 28.94 29.82 19.53 0 111.5 13.02 29.79 342.4
2011-08-26 00:01:00 NaN NaN NaN NaN NaN NaN NaN NaN
2011-08-26 00:01:30 NaN NaN NaN NaN NaN NaN NaN NaN
2011-08-26 00:02:00 NaN NaN NaN NaN NaN NaN NaN NaN
print df.resample('30S', fill_method='ffill').head()
1 2 3 4 5 6 7 \
0
2011-08-26 00:00:00 1155179 3.232 23.70 3.281 0.386 25.27 111.5665
2011-08-26 00:00:30 1155180 3.289 20.44 2.153 0.222 25.25 111.5735
2011-08-26 00:01:00 1155180 3.289 20.44 2.153 0.222 25.25 111.5735
2011-08-26 00:01:30 1155180 3.289 20.44 2.153 0.222 25.25 111.5735
2011-08-26 00:02:00 1155180 3.289 20.44 2.153 0.222 25.25 111.5735
8 9 10 11 12 13 14 15
0
2011-08-26 00:00:00 28.92 29.83 19.13 0 111.5 13.02 29.77 345.7
2011-08-26 00:00:30 28.94 29.82 19.53 0 111.5 13.02 29.79 342.4
2011-08-26 00:01:00 28.94 29.82 19.53 0 111.5 13.02 29.79 342.4
2011-08-26 00:01:30 28.94 29.82 19.53 0 111.5 13.02 29.79 342.4
2011-08-26 00:02:00 28.94 29.82 19.53 0 111.5 13.02 29.79 342.4
关于python - 使用 pandas.date_range 和 pandas.reindex python 填充时间序列数据中的缺失点，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/35739208/