熊貓減去兩組數據

Question

我想弄清楚爲什麼如果我減去兩組數據，它會給我NaN錯誤。我想做一個元素乘元素減法，我已經成功地提取了兩套獨立的，但是當我嘗試用它做什麼，它不斷給我NaN的錯誤

這裏是我的代碼：

import pandas as pd 
import numpy as np 
import matplotlib.pyplot as plt 

filename = 'vc_test.csv' 
headings = 'injection', 'treatment','filename', 'l0', 'l1', 'l2', 'l3', 'filename', 'r0', 'r1', 'r2', 'r3' 

#import data 
df = pd.read_csv(filename, names=headings) 




#convert values to floats 
df['l0'] = df['l0'].astype(float) 
df['l1'] = df['l1'].astype(float) 
df['l2'] = df['l2'].astype(float) 
df['l3'] = df['l3'].astype(float) 
df['r0'] = df['r0'].astype(float) 
df['r1'] = df['r1'].astype(float) 
df['r2'] = df['r2'].astype(float) 
df['r3'] = df['r3'].astype(float) 

#group by injection so you have 2 subgroups 
df.groupby(['injection']) 

#subtract and assign to a different array to normalize data to baseline 
df_os, df_od = df[['l1','l2','l3']].sub(df['l0'], axis=0), df[['r1','r2','r3']].sub(df['r0'], axis=0) 

print df_os 
print df_od 

ioc = df_os - df_od 
print ioc 

我也試過df.sub（），但沒有奏效。請幫助...

Answer 1

列名稱存在問題 - 在兩個減法文件DataFrames中都需要相同的列名稱。也可用於創建DataFrames，您可以使用boolean indexing。

import pandas as pd 
import io 


temp=u""" 
1,a1453905960,a,95.4500,95.0900,95.0980,433810,s,95.4500,95.0900,95.0980,433810 
1,a1453906020,b,95.4700,94.9500,95.4500,934980,d,85.4520,94.100,95.7980,433810 
2,a1453906080,f,95.1000,94.8700,95.0900,791657,e,95.4500,934980,9.4400,85.4520 
2,a1453906140,c,4.0300,94.7000,94.9620,763531,w,95.1000,94.8700,95.0900,791657 
2,a1453906200,f,95.0300,94.8200,94.8918,501298,r,95.1000,94.8700,95.0900,791657""" 

headings =[ 'injection', 'treatment','filename', 'l0', 'l1', 'l2', 'l3', 'filename', 'r0', 'r1', 'r2', 'r3'] 

#after testing replace io.StringIO(temp) to filename 
df = pd.read_csv(io.StringIO(temp), names=headings) 
print df 
    injection treatment filename  l0  l1  l2  l3 filename \ 
0   1 a1453905960  s 95.45 95.09 95.0980 433810  s 
1   1 a1453906020  d 95.47 94.95 95.4500 934980  d 
2   2 a1453906080  e 95.10 94.87 95.0900 791657  e 
3   2 a1453906140  w 4.03 94.70 94.9620 763531  w 
4   2 a1453906200  r 95.03 94.82 94.8918 501298  r 

     r0   r1  r2   r3 
0 95.450  95.09 95.098 433810.000 
1 85.452  94.10 95.798 433810.000 
2 95.450 934980.00 9.440  85.452 
3 95.100  94.87 95.090 791657.000 
4 95.100  94.87 95.090 791657.000 

#convert values to floats 
df[['l0','l1','l2','l3','r0','r1','r2','r3']] = df[['l0','l1','l2','l3','r0','r1','r2','r3']].astype(float) 


#select data by column injection to 2 new dataframes by boolean indexing 
df_os = df[df['injection'] == 1] 
#print df_os 

df_od = df[df['injection'] == 2] 
#print df_od 

#subtract and assign to a different array to normalize data to baseline 
df_os, df_od = df[['l1','l2','l3']].sub(df['l0'], axis=0), df[['r1','r2','r3']].sub(df['r0'], axis=0) 

print df_os 
     l1  l2   l3 
0 -0.36 -0.3520 433714.55 
1 -0.52 -0.0200 934884.53 
2 -0.23 -0.0100 791561.90 
3 90.67 90.9320 763526.97 
4 -0.21 -0.1382 501202.97 

print df_od 
      r1  r2   r3 
0  -0.360 -0.352 433714.550 
1  8.648 10.346 433724.548 
2 934884.550 -86.010  -9.998 
3  -0.230 -0.010 791561.900 
4  -0.230 -0.010 791561.900 

#set column names in df_od by df_os 
df_od.columns = df_os.columns 
ioc = df_os - df_od 
print ioc 
      l1  l2   l3 
0  0.000 0.0000  0.000 
1  -9.168 -10.3660 501159.982 
2 -934884.780 86.0000 791571.898 
3  90.900 90.9420 -28034.930 
4  0.020 -0.1282 -290358.930 

ioc = df_os.sub(df_od) 
print ioc 
      l1  l2   l3 
0  0.000 0.0000  0.000 
1  -9.168 -10.3660 501159.982 
2 -934884.780 86.0000 791571.898 
3  90.900 90.9420 -28034.930 
4  0.020 -0.1282 -290358.930 

或者您也可以通過values轉換DataFrames到numpy arrays然後。減去：

ioc = df_os.values - df_od.values 
print ioc 
[[ 0.00000000e+00 0.00000000e+00 0.00000000e+00] 
[ -9.16800000e+00 -1.03660000e+01 5.01159982e+05] 
[ -9.34884780e+05 8.60000000e+01 7.91571898e+05] 
[ 9.09000000e+01 9.09420000e+01 -2.80349300e+04] 
[ 2.00000000e-02 -1.28200000e-01 -2.90358930e+05]] 

print pd.DataFrame(ioc, columns=['a','b','c']) 
      a  b   c 
0  0.000 0.0000  0.000 
1  -9.168 -10.3660 501159.982 
2 -934884.780 86.0000 791571.898 
3  90.900 90.9420 -28034.930 
4  0.020 -0.1282 -290358.930