Warning: Integer operands are required for colon operator when used as index
5 visualizaciones (últimos 30 días)
Mostrar comentarios más antiguos
Matthias Pospiech
el 20 de Feb. de 2012
Comentada: Walter Roberson
el 8 de Dic. de 2020
I am using fft functions and have twice in my code this warning message, bu no clue why I get the warning:
N = 2048*16;
sizeData = size(data,2);
padding = (N-sizeData)/2;
cInput = zeros([1 N]); % <-- Warning
cInput(padding+1:(padding+sizeData)) = data;
cInput = fftshift(cInput);
% transform
cfourier = fftshift(fft(cInput));
% iFFt
cOutput = ifft(ifftshift(cfourier));
cOutput = ifftshift(cOutput); % <-- Warning
AmpIFourier = abs(cOutput(padding+1:(padding+sizeData)));
1 comentario
Walter Roberson
el 8 de Dic. de 2020
Haldun Hannis posted a question in the Answer section but deleted it while I was composing the reply. That reply follows:
(NFFT was constructed with 2^nextpow2 so is a power of 2)
WINDOW = hamming(NFFT/8+1); % Windowing function
NOVERLAP=length(WINDOW)*0.5;
provided that the data is not very small, NFFT will be exactly divisible by 8 so NFFT/8 would be an even integer unless NFFT is 8 or less (not impossible, should be tested for.) With NFFT/8 being even, NFFT/8+1 will be odd. So you are asking for an odd-size hamming filter to be created.
You then take the length of that, which will be odd. And you multiply by 0.5. With it being odd, you will not get an integer result.
You then pass that non-integer value as the overlap size, into a function that uses it something vaguely like
data(start:start+overlap-1)
and that is a problem because overlap is not an integer
You should be using floor() or ceil() to force integer
Respuesta aceptada
Image Analyst
el 20 de Feb. de 2012
(N-sizeData)/2 can be a fractional number, e.g. "something point 5" so try
padding = floor((N-sizeData)/2);
or better yet, just use the padarray() function to do your padding.
5 comentarios
Jan
el 21 de Feb. de 2012
@Matthias: I've never seen a bug in Matlab, which leads to a wrong line number for warnings. While I assume IA and Oleg point to the correct solution, there must be an additional problem.
Hannetjie
el 6 de Mzo. de 2015
@Jan Simon: It is now 3 years later so my apologies if this has since been touched upon in newer posts... But I came across this same anomaly (and this is the first and only time I've seen it). When running my code through, I get
Warning: Integer operands are required for colon operator when used as index > In tdnoise3 at 73
The warning in my case makes no sense at line nr 73. However, if I step through the exact same code, I get
Warning: Integer operands are required for colon operator when used as index > In tdnoise3 at 81
Which does make sense and was where the fix was required.
But OK, has no bearing on the topic of this post.
Más respuestas (2)
Oleg Komarov
el 20 de Feb. de 2012
Perfectly fine
idx = 2.5:9;
But when you use to index a position:
A = rand(10,1);
A(idx)
There's no such a thing as position number 2.5
0 comentarios
Jan
el 20 de Feb. de 2012
N = 2048*16;
cInput = zeros([1 N]); % <-- Warning
No, I do not get a warning in this code. Are you sure, that this is the original code?
0 comentarios
Ver también
Categorías
Más información sobre Entering Commands en Help Center y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!