如何使用python将具有不同列的TXT文件转换为带有定界符“，”的文件a？

Question

I have the following ORIGINAL .TXT file:

"TRACE: A"
"FORMAT: WOW"

"Frequency" "Data 1"    "Data 2"
1   1   2
1   6   0

"Frequency" "Data 1"    "Data 2"
1   5   0
1   6   0

In order to run python script, I have to first open .TXT file in LibreOffice, then save and close it, to create EXPECTED .TXT file with a delimiter=',', as shown below:

"TRACE: A",,
"FORMAT: WOW",,
,,
"Frequency","Data 1","Data 2"
1,5,0
1,5,0

"Frequency","Data 1","Data Trace Imag"
1,5,0
1,5,0

Now I can use the following code to plot the graphs:

import numpy as np
from pylab import *

fileName = 'EXPECTED.TXT';

traceAdata= np.genfromtxt(fileName, delimiter=',')[3:5].T;

x = traceAdata[0]*1e-6;
A = traceAdata[1];

plot(x,A);
savefig(fileName + '.png');

I have tried to use different delimiters in the original TXT file such as: ' ', None, '\\t', but none of them worked, and I had following error:

Line #5 (got 3 columns instead of 1)

Is there any other way to convert original TXT file to the file with needed delimiter using python?

UPD: Current Solution, thank you @lxop.

import numpy as np
from pylab import *

fileName = 'C2.TXT';

traceAdata= np.genfromtxt(fileName, skip_header=21,skip_footer = 334).T;

x = traceAdata[0]*1e-6;
A = traceAdata[1];

plot(x,A);
savefig('traceAdata.png');

traceBdata= np.genfromtxt(fileName, skip_header=337).T;
x = traceBdata[0]*1e-6;
B = traceBdata[1];

plot(x,B);
savefig('traceBdata.png');

Answer 1

You can load the space separated file directly, without converting, with a script like

import numpy as np
from pylab import *

fileName = '11.TXT';

traceAdata= np.genfromtxt(fileName, skip_header=4).T;

x = traceAdata[0]*1e-6;
A = traceAdata[1];

plot(x,A);
savefig(fileName + '.png');

In particular not the use of skip_header to skip over the lines that don't have data. Also, if you explicitly specify delimiter=' ' then genfromtxt will treat each space as a delimiter and try to get values between the individual space characters, so just leave that argument out.

If you really want to convert it to a CSV first, then you should look at a parsing library, since it will need to handle quoting etc.