matlab - Linearity of filtering system on concatenated signals -

i have signal s[n] , bandpass filter filters out low , high frequency components of signal. want store signal matlab array , put through filter.

however, cannot store s[n] in matlab because allocated memory not sufficient contain such long signal. decided split signal n segments, , put each segment through bandpass filter, , lastly assemble them after filtering.

i wondering if there linearity issue approach. if not valid, there other method can achieve want? thanks.

if using fir filter can using convolution:

x = rand(1000,1) b = fir1(100, 0.5) y1 = zeros(1100, 1) % compute response using first 400 points of x y1(1:500) = conv(x(1:400),b) % compute response using last 600 points of x y1(401:1100) = y1(401:1100) + conv(x(401:1000),b) % compute whole response compare y2 = conv(x,b) 

note length of conv(a,b) length(a) + lenght(b) - 1

but total response have same length of approach, going have same memory problem.