Loop for removing values from a column vector

Hi all!

So I have a column vector that I imported from Excel spreadsheet with a certain amount of values (3000 something). What I'm trying to do is create a script with a loop that will go through every value in the vector and if the value of element k is <= than the value of element k-1, it deletes this value and goes to the next one.

Example: column_vector = [1 2 1 3 7 7 5 4 8 6 9 2] -------------> new_column_vector = [1 2 3 7 8 9]

Basicaly I want it to go up withou oscilating. Any help is appreciated (I havent touched Matlab for at least 5 years)

[Nikita Kaminskyy]

An important remark: my column_vector has values with several decimal places (10 or so), they're all double values

Answer 1

[Cris LaPierre]

I would do this with logical indexing.

column_vector = [1 2 1 3 7 7 5 4 8 6 9 2];
logInd = [1 diff(column_vector)]>0; % Since diff will have n-1 values, prepend with 1
new_column_vector = column_vector(logInd)

new_column_vector = 1×6

1 2 3 7 8 9

●

[Nikita Kaminskyy]

I'm getting this error:

Error using horzcat

Dimensions of arrays being concatenated are not consistent.

[Cris LaPierre]

It works for the example you've provided, as you can see.

The error sounds like your data is a column vector instead of a row vector. In that case, make logInd a column vector as well (use [1; ...] instead of [1 ...]).

column_vector = [1 2 1 3 7 7 5 4 8 6 9 2]';
logInd = [1; diff(column_vector)]>0; % Since diff will have n-1 values, prepend with 1
new_column_vector = column_vector(logInd)

new_column_vector = 6×1

1 2 3 7 8 9

●

[Nikita Kaminskyy]

Is it only for integers? It only seems to work until a certain part of the vector but it just doesn't what it's supposed to :/

[Cris LaPierre]

It is not specific to integers, but it only looks at immediate neighbors. That could lead to a scenario where the final vector in not always strictly increasing.

I think a better solution would use cummax. This will give you the running max. I use unique to removed duplicates.

column_vector = rand(10,1)*10

column_vector = 10×1

2.1816 8.4985 1.3067 0.5397 3.1610 9.1202 7.1885 7.6020 9.8508 7.0116

●

new_column_vector = unique(cummax(column_vector))

new_column_vector = 4×1

2.1816 8.4985 9.1202 9.8508

●

[Nikita Kaminskyy]

I guess there must have been a misunderstanding of the problem. Your code is calculating the difference from two consecutive values and in fact it works for the given example. Imagine I have the following vector:

v = [200.880438000000

201.080292400000

201.264694400000

201.461538600000

197.282910400000

197.743862800000

198.381722000000

198.644310200000

198.974400800000

199.502214800000

200.040847000000

200.149417000000

200.473740600000

201.053679000000

201.173782400000

201.330348800000]

The output I'm expecting is:

new_v = [200.880438000000

201.080292400000

201.264694400000

201.461538600000

201.583782400000

201.690348800000]

[Nikita Kaminskyy]

cummax worked like a charm, thank you!

Answer 2

[Fangjun Jiang]

%%
v=[1 2 1 3 7 7 5 4 8 6 9 2];
for k=length(v):-1:2
    if v(k)<=v(k-1)
        v(k)=[];
    end
end

[Nikita Kaminskyy]

Hi, thanks for a quick answer but it doesn't work