Error in defining x values for peak analysis in Matlab

Hi. I have a problem in creating plot for peak analysis. It can plot the data but not for the x-axis. My data is in the form of double, and my x-axis is datetime. However, when I plot findpeaks, I got an error warning :

"Error using findpeaks
Expected X to be increasing valued.

Error in findpeaks>parse_inputs (line 242)
      validateattributes(seconds(Xin-Xin(1)),{'double'},{'real','finite','vector','increasing'},'findpeaks','X');

Error in findpeaks (line 136)
  = parse_inputs(isInMATLAB,Yin,varargin{:});"

The value of my datetime data actually increase by the number of rows, so I wonder why I got an error. I attach the picture of my datetime data. Could you please help me? I have been working on this problem since morning and up to know I got nothing :(. Thank you for your help.

[Ganavi Mg]

Hi Even I got same error as you got. My code is clc; clear all; load('DATA_01_TYPE01.mat'); [n, p] = size(sig); t=1:n; plot((1:1000),sig(2,1:1000)); xlabel('samples') ylabel('sample amplitude') title('User1') figure; n=1000; ts =0.0001; ws = 2*pi/ts; f = fft(sig(2,1:1000)); w = ws*(-n/2:(n/2)-1)/n; plot(w,abs(f)) xlabel('frequency'); ylabel('amplitude'); title('fast fourier transform of PPG DATA'); [pks,locs]= findpeaks(frequency,sample amplitude); In the place of frequency,sample amplitude what value we must use. I need peak analysis of signal. The data is in .mat format. I had loaded the mat values in to my code, then found out the fft of the data. I need peak of the signals. What I should do now ?

Answer 1

[Walter Roberson]

I suspect you are using R2015b or earlier; datetime was not permitted for the "location" parameter until R2016a.

[Kasih Ditaningtyas Sari Pratiwi]

No I am using R2017a so it should work. My datetime also has increasing value up to the last row. Or do you think it is because the date I use remains the same for some variables, for example :

03.10.2017 00:00:15
03.10.2017 00:00:40

but still they are increasing. I am so confused about this :(

[Walter Roberson]

Supposing t is your datetime variable, then look at

temp = minutes(diff(t));
badpoints = find(temp <= 0);
t([badpoints(:), badpoints(:)+1])

[Kasih Ditaningtyas Sari Pratiwi]

Walter, you are right. Some datetime are duplicates. Do you know how to delete duplicate based on duplicate datetime, and at the same time delete another data in their rows? Thank you so much for your help

[Walter Roberson]

tisdup = [false; diff(t(:)) == 0];
t(tisdup) = [];
data(tisdup,:) = [];

[Kasih Ditaningtyas Sari Pratiwi]

Thank you for your reply, Walter. I got an error :Deletion requires an existing variable." Here is the code I input :

%%Peak Flow Analysis
% analyze badpoints for datetime data before it is used as x-axis. This is 
% due to the requirement of increasing value for x-axis. Badpoints are the 
% data which duplicates. If there are badpoints datetime data can not be 
% used as x-axis.
t=finalCSVnew{:,1};
temp = minutes(diff(t));
badpoints = find(temp <= 0);
t([badpoints(:), badpoints(:)+1]);

% delete all data in the rows where there is duplicate in datetime data
tisdup = [false; diff(t(:)) == 0];
t(tisdup) = [];
data(tisdup,:) = [];

Do you think I have to mention the position of another column? I have 5531x15 table there in my data (finalCSVnew).

[Walter Roberson]

You do not need most of that.

t = finalCSVnew{:,1};
tisdup = [false; diff(t(:)) == 0];
finalCSVnew(tisdup, :) = [];

[Kasih Ditaningtyas Sari Pratiwi]

Awesome! Thank you so much Walter! You're really smart! Hopefully when later I am no longer a newbie in matlab programming, I can answer the question as you do. Thank you again :D