How can I insert missing rows in a matrix

Hi Guys I have a matrix like this: x=[1 4; 4 6;2 3;3 2;4 2;1 3]. where I am showing data of three days. The first row is like a serial number and second row contains the data. I should have 4 data in each day and the first row of the matrix should look like this: x(:,1)= (1 2 3 4 1 2 3 4 1 2 3 4)' But as I have missing data i don't have this matrix. Now I want to create a matrix where all the missing data will be shown with NaN or zero in the second column. So the output will look like : x=[1 4; 2 0;3 0;4 6;1 0;2 3;3 2;4 2;1 3;2 0;3 0;4 0]. That means I will show all the serial numbers in each day and 0 for the missing data. Can anyone help plz? Thank you.

Answer 1

[Jos (10584)]

This is not really trivial. Here is one approach:

x=[1 4; 4 6;2 3;3 2;4 2;1 3]

Data = x(:,2) ;
SerialNumber = x(:,1) ;
StartNewDay = diff([Inf ; SerialNumber]) < 0
Day = cumsum(StartNewDay)

NewX(:,1) = repmat(1:max(SerialNumber),1,max(Day)).'
M = accumarray([SerialNumber Day],x(:,2))  % build a matrix with zeros
NewX(:,2) = reshape(M,[],1)

This assumes that the serial numbers are consecutive positive integers.

[Mohammad Sayeed]

It worked perfectly Brother. Thank you very much. Yes the serial numbers are consecutive positive integers. But say, if I suppose to have consecutive time periods in the first row such as 12:01:00 AM and so on instead of consecutive numbers then is it possible solve the problem if I have missing data in some time periods? Please Let me if you can Help me. However, thanks again for the above codes.

[Mohammad Sayeed]

Dear @Jos, If I have more observations per day say I will have hourly data so I have serial number 1:24 in each day then how can I make required changes in your codes as I have different number of observations in different data set? Please let me know.