splitting data in a table when ever the cell is empty

Hi

I have a long set of data where 100 cycles appear in one colum under each other. Each cycle varies in length.

However they are split up with two empty cells etween each data set.

I want to split the data so the 100 cycles appear next to each other so instead of one colum with all the data there is 100. ( not necerssary but if possible ideally i would want to create a 'cell' numbered 1-100 in first colum with the data in the second colum.)

The following is just an example of what i want.3 cycles side by side.

x  
10 20
20 6
30 8
40 7


50 5
40 4
20 6


45 0
4  3
30 6
8  8


10  20 50 5 45 0
20  6  40 4 4  3
30  8  20 6 30 6
40  7       8  8

[Guillaume]

Since you mention empty cells, is your data stored in a cell array?

It's much easier if you use valid matlab syntax to describe your data. That way there's no ambiguity. Or attach an example of the data as a mat file.

Other than the empty cells, is all the data numeric? If so, it will be much more efficient to store the final data in a matrix with NaN in the final rows of the shorter cycles. A cell array of scalar numeric uses 15 times more memory than a matrix with the same numbers, and is significantly more difficult to use.

[Guillaume]

Sufia Fatima's comment mistakenly posted as an answer moved here:

hi

I have imported the data from a .txt file and yes all the data is numerical.

e above example is what the data looks likeTh.

[Guillaume]

Please, either attach demo text file and the code you used to import it, or attach your data as a mat file.

It's really unclear what form your data is in matlab. You haven't answered my question: is your data stored in a cell array?.

[Sufia Fatima]

hi i have attched the .mat file.

kind regards

[Guillaume]

So your data is stored in a matrix. Matrices don't have empty cells. I'm confused by your question.

[Sufia Fatima]

There are defo empty rows. for example 15811 and 17030.

wherever there is an empty row i want to split the data to signify a diff test.

Answer 1

[Guillaume]

Matrices can never have empty rows. What you call empty rows are rows with 0s in both columns.

If I understood correctly what you want to do:

%indentation_3hr: input matrix
zerorows = find(all(indentation_3hr == 0, 2));  %find rows with 0 in both columns
rowdist = diff([0; zerorows; size(indentation_3hr, 1)]);  %distance between the rows containing 0
splitted = mat2cell(indentation_3hr, rowdist, size(indentation_3hr, 2));  %split into cell array at the 0s
splitted = cellfun(@(m) m(1:end-1, :), splitted, 'UniformOutput', false); %remove last row of each matrix (as it's [0 0])
splitted(cellfun(@isempty, splitted)) = [];  %remove empty matrices (due to the double rows of [0 0])
maxheight = max(cellfun(@(m) size(m, 1), splitted));  %get height of the tallest matrix
padded = cellfun(@(m) [m; nan(maxheight - size(m, 1), size(m, 2))], splitted, 'UniformOutput', false);  %pad shorter matrices with nan

newmatrix = [padded{:}];  %concatenate all padded matrices horizontally

newmatrix has 200 columns (2 columns per cycle). Shorter cycles are padded with NaN (since again, matrices cannot have empty rows).

[Sufia Fatima]

Thank You Guillaume this worked.

Answer 2

[Walter Roberson]

mask = ~any(indentation_3hr, 2).';  %want row vector
starts = strfind([false mask],[0 1]);  %start of non-empty groups
stops = strfind([mask false], [1 0]);  %end of non-empty groups
extracted = arrayfun(@(FROM, TO) indentation_3hr(FROM:TO, :), starts, stops, 'uniform', 0);
maxlen = max(arrayfun(@(V) size(V,1), extracted));

FirstN = @(M,N) = M(1:N,:);
PadN = @(M,N) FirstN([M;nan(N, size(M,2))], N);

result = cell2mat( cellfun(@(M) PadN(M, maxlen), extracted(:).', 'uniform', 0) );

[Sufia Fatima]

Hi Thank you for your help. However, I have this error when i use this code

Error: File: indentation.m Line: 15 Column: 17

Incorrect use of '=' operator. To assign a value to a variable, use '='. To compare values for equality, use '=='.

This is in relation to

FirstN = @(M,N) = M(1:N,:);

I have done what was suggested with inputting

FirstN = @(M,N) == M(1:N,:);

However, i still get the error

Error: File: indentation.m Line: 15 Column: 17

Invalid use of operator.

Again Thanks for the suggestion

[Walter Roberson]

FirstN = @(M,N) M(1:N,:);