How can I insert rows of NaN into a vector of discontinuous values

Hey guys

Working with some data which have a ton of gaps; what I want to do is expand the data matrix into the size of a reference vector, filling NaN's along the way.

The first column in my data matrix represents dates. I want to insert rows of NaN's where the dates are missing. So if datavec represents the first column in my data and refvec represents what it should look like, I would like the output matrix' first column to look like newvec.

datavec = [1992 1994 1995 1996 2000] refvec = [1992 1993 1994 1995 1996 1997 1998 1999 2000]

newvec = [1992 NaN 1994 1995 1996 NaN NaN NaN 2000]

I've tried combinations of the insertrows.m file on the exchange and a few other options, all of which failed miserably

Cheers!

Answer 1

[Azzi Abdelmalek]

datavec = [1992 1994 1995 1996 2000]
refvec  = [1992 1993 1994 1995 1996 1997 1998 1999 2000]
newvec = nan(size(refvec))
newvec(ismember(refvec,datavec))=datavec

[Aakash]

Hey Azzi

Good answer, however when I tried to adapt to my particular situation it doesn't seem to go through. Let me clarify; what I have is a matrix of data. The first column represents the dates (i.e. datavec in your example). I want to then insert rows of NaN, not just replace the values with NaN as I think your example effectively does. Any thoughts?

[Azzi Abdelmalek]

datavec = [1992 1994 1995 1996 2000]
refvec= datavec(1):datavec(end)
idx=~ismember(refvec,datavec)
newvec = nan(size(refvec))
newvec(ismember(refvec,datavec))=datavec

[Aakash]

Close, let me try and explicate more.

I have a matrix: the rows in Column 1 represents the dates and the rows in column 2:end is the actual data corresponding with each date.

What I would like to do is go through each element of Column 1, which are discontinuous dates and insert rows of NaNs

reference vec = [ 1992:2000]

datamat = [ 1992 1 2 3 4 5 6 7 8 9....; 1994 1 2 3 4 5 6 7 8 9...; ...]

newmat = [1992 1 2 3 4 5 6 7 8 9 ; NaN NaN NaN NaN NaN NaN; 1994 1 2 3 ...]

Does this make sense?

[Azzi Abdelmalek]

refvec =(1992:2000)'
datamat = [ 1992 1 2;1994 5 1;2000 5 8]
newmat=nan(numel(refvec),size(datamat,2))
newmat(ismember(refvec,datamat(:,1)),:)=datamat

[Aakash]

Awesome, thanks!

The intuition as I see it is, fill a matrix with NaN, to be replaced later on. We then find where our actual dates match our reference dates. Then replace those values, leaving you with nothing but NaNs.

[anas ibnu]

hi i have a similar issue as yours .. and i tried the above (lastone) code and it gives an error "Subscripted assignment dimension mismatch."

can you please say what does it means .thanks in advance

Answer 2

[Alfredo guerrero]

Hi Guys I am new in Matlab.

I have two different variables in size, how can i fill with NaN var2 in order to have the same length I mean I need the result in var2 = [2;4;3;2;1;4;5;4;NaN;NaN;NaN;NaN;NaN;NaN;NaN]

%% Real DATA

var1 = [1;2;3;4;5;6;7;8;9;10;11;12;13;14;15];

var2 = [2;4;3;2;1;4;5;4]

I hope someone help me!