将输出的多维一维数组存储为数据文件

Question

import numpy as np 
import matplotlib.pyplot as plt 
x = open(r'''C:\Users\Documents\ex.txt''')

[INPUT ex.txt file:
   -1.642195902 0.751055263
    0.496998351 -1.306558434
   -0.490237525 -0.188855324
   -1.357284374 0.282238191
   -0.160982328 -1.115393803
    1.167022948 0.564800286
    -2.050084963    0.262537079
    0.883449292 -0.276800002
    1.017703957 -0.044710318
    0.194025051 -1.392772391
    0.209566571 -0.314937244
    1.840331474 -1.544109096
    0.493878872 -0.405593557


   ]

ls = x.readlines()

x1 = np.array([])
x2 = np.array([])
for l in ls: 

    col = l.split()

    x1 = np.append(x1,float(col[0])) 
    x2 = np.append(x2,float(col[1]))
    x3 = np.polyfit(x1,x2,1)
    print(x3)


 Output is:

 [-0.22867408  0.37552763]
 [-0.96186389 -0.82851367]
 [-0.95783751 -0.77027985]
 [-0.90374253 -0.79169204]
 [-0.97569702 -0.93091431]
 [-0.29008702 -0.26483726]
 [-0.27811928 -0.26755408]

     .........

 [-0.06747052 -0.0218189 ]

My goal is to use the values for x3 and subtract from the original values x1 and x2, I was thinking in saving the output as a .txt file but the output is several 1 dimensional arrays.

Does anyone know how could I save the result from x3 such that I would be able to substract it from my original values? I am not sure how to code it. My goal is to detrend a set of data.

Many thanks

Answer 1

Try this:

import numpy as np 
import matplotlib.pyplot as plt 
x = open(r'''C:\Users\Documents\ex.txt''')

ls = x.readlines()

x1 = np.array([])
x2 = np.array([])
x_array = np.array([])
x3_array = np.array([])

for l in ls: 

    col = l.split()
    x_array = np.append(x_array, [float(col[0]), float(col[1])])
    x1 = np.append(x1,float(col[0])) 
    x2 = np.append(x2,float(col[1]))
    x3_array = np.append(x3_array, np.polyfit(x1,x2,1))

sub_res = x_array - x3_array
print (sub_res)

Answer 2

first regarding importing your data:

data = np.genfromtxt('c:/Users/kaufnbnj/Desktop/ex.txt', names = ['x1', 'x2'])

array([(-1.6421959 ,  0.75105526), ( 0.49699835, -1.30655843),
       (-0.49023753, -0.18885532), (-1.35728437,  0.28223819),
       (-0.16098233, -1.1153938 ), ( 1.16702295,  0.56480029),
       (-2.05008496,  0.26253708), ( 0.88344929, -0.2768    ),
       ( 1.01770396, -0.04471032), ( 0.19402505, -1.39277239),
       ( 0.20956657, -0.31493724), ( 1.84033147, -1.5441091 ),
       ( 0.49387887, -0.40559356)],
      dtype=[('x1', '<f8'), ('x2', '<f8')])

accessing first column:

data['x1']
array([-1.6421959 ,  0.49699835, -0.49023753, -1.35728437, -0.16098233,
        1.16702295, -2.05008496,  0.88344929,  1.01770396,  0.19402505,
        0.20956657,  1.84033147,  0.49387887])

Then the fitting part:

x3 = np.polyfit(data['x1'], data['x2'], 1)

Note that polyfit Returns the coefficients of your Fitting function, not a sampled array for your input data:

x3
array([-0.33019369, -0.3484815 ])

So you can plot input data and fit in a plot like this:

x_fit = x3[0]*data['x1'] + x3[1]
plt.plot(data['x1'], data['x2'], 'o')
plt.plot(data['x1'], x_fit)

... and x2 after detrend would be:

x_detrend = data['x2'] - x_fit
plt.plot(data['x1'], x_detrend, 'o')