FFT filtering noise?

Why I did I get a final signal like the first signal after denoising??

filename = 'koo.CSV';
M = csvread(filename);
x=M(:,3);
Y=M(:,4);
fs=250;
f=10;

subplot(3,2,1)
plot (x,Y,'b','linewidth',2)
title('Original Arc Signal')
xlabel('TIME,t(s)')
ylabel('VOLTAGE,V');

yrand = (rand(size(Y))-0.5) * 10; 
z=Y + yrand;
subplot(3,2,2)
plot (x, z,'b', 'linewidth', 2)
title('Noisy Arc Signal')
xlabel('TIME,t(s)')
ylabel('VOLTAGE,V');

L=length(z);
NEFT=2^nextpow2(L);
z_fft=abs(fft(z,NEFT));
freq=fs/2*linspace(0,1.5,NEFT/2+1);
subplot(3,2,4);
plot(freq,z_fft(1:length(freq)));

o=5;
wn=[1 10]*2/fs;
[b,a]=butter(o,wn,'bandpass');
figure;freqz(b,a,1000,fs);
figure(1)
[h,w]=freqz(b,a,1000,fs);
subplot(3,2,3);
plot(w,20*log10(abs(h)));

z_filt=filter(b,a,z);
subplot(3,2,5);
ylim([0 20])

plot(x,z_filt,'b','linewidth', 2);

[Image Analyst]

You forgot to attach koo.csv.

What exact line of code is the filter actually being applied to the noisy version of the original signal?

[hafis radzi]

hello sir,i already upload the file and add missing coding.i hope u can help me

[Image Analyst]

It looks roughly like it if you invert the signal - I think it's just upside down.

[hafis radzi]

i already tranpose the polatry of data.figure below,but not same with the original signal.i just want use FFT tecnique to denoise the origal noise signal.