How to replace all the values of a variable in a dataset following a condition

Hi all. I have a dataset 'data' (16 x 70 fields), which is a structure. Each field has 14 vectors, 1 with strings (1 value) and the rest with integers (1 value per field). I want to replace all the values of one of those vectors for each field (let's call it VECTOR) for the whole data set following the criteria: if it's 0, replace it with 1, if it's 1, replace it with 0. I've tried this:

data(data.VECTOR == 1) = 0;
data(data.VECTOR == 0) = 1;

But it doesn't work. Any ideas?

By the way, there are only three possible values: -1, 0, 1. I want to replace the 0 with 1, and the 1 with 0, while leaving the -1 unchanged.

[madhan ravi]

can you attach your dataset as a .mat file?

[Renzo Lanfranco]

Sure. The fields I have to change are sideReport and sideAccu. If they are 0, they have to be replaced to 1, and vice versa.

[Walter Roberson]

data.VECTOR = ~data.VECTOR;

[Renzo Lanfranco]

It says "Too many input arguments"

[Walter Roberson]

for idx = 1 : numel(data)
    data(idx).VECTOR = ~data(idx).VECTOR;
end

[Renzo Lanfranco]

Thanks. It says "reference to non-existent field". I don't understand why does it say that. Every field has 14 vectors, one of which is VECTOR.

[Walter Roberson]

for idx = 1 : numel(data)
    data(idx).sideReport = ~data(idx).sideReport;
    data(idx).sideAccu = ~data(idx).sideAccu;
end

[Renzo Lanfranco]

Thanks so much, Walter. One quick question: Those vectors can only have 3 values: -1, 0, and 1. Does your for look leave the -1 values unchanged?

[Renzo Lanfranco]

I tried the code, and it turns -1 values to 0. How can you turn the 0 into 1 and the 1 into 9, but leave the -1 unchanged?

[Walter Roberson]

for idx = 1 : numel(data)
    mask = data(idx).sideReport == 1;
    data(idx).sideReport(mask) = 9;
    mask = data(idx).sideReport == 0;
    data(idx).sideReport(mask) = 1;
    mask = data(idx).sideAccu == 1;
    data(idx).sideAccu(mask) = 9;
    mask = data(idx).sideAccu == 0;
    data(idx).sideAccu = 1;
end

Order is important. You were changing all 1 to 0 and then changing all 0 including the ones you just put there to 1.

If you needed to switch 0 to 1 and 1 to 0 without affecting anything else you could do something like

mask0 = data(idx).sideReport == 0;
mask1 = data(idx).sideReport == 1;
data(idx).sideReport(mask0) = 1;
data(idx).sideReport(mask1) = 0;

[Walter Roberson]

Another approach if your original values are only -1, 0, 1:

map = [-1, 1, 9];    %new values for -1, 0, 1
for idx = 1 : numel(data)
    data(idx).sideReport = map(data(idx).sideReport + 2);
    data(idx).sideAccu = map(data(idx).sideAccu + 2);
end

adding 2 transforms the -1, 0, 1 (not valid indices) to [1 2 3], which can be used in a lookup table.

[Renzo Lanfranco]

Thanks so much, Walter. This solution worked perfectly:

mask0 = data(idx).sideReport == 0;
mask1 = data(idx).sideReport == 1;
data(idx).sideReport(mask0) = 1;
data(idx).sideReport(mask1) = 0;