views:

75

answers:

2

This question is related to http://stackoverflow.com/questions/3071558/matlab-how-to-merge-data

how to use the FIND function with IF statement? e.g. if I have this data:

20  10  1
20  11  1
20  15  1
23  10  1
23  10  1
23  12  0

Rule 1: Data of column 3 must be 1.

Rule 2: If n is the current index of column 1, if n equal n-1 (20=20) of column 1, data of column 2 of index n and n-1 is merged.

20  21  0
20  15  0
20  0   0
23  20  0
23  0   0
23  12  0

EDITED:

fid=fopen('data.txt');
A=textscan(fid,'%f%f%f');
fclose(fid);
in = cell2mat(A)'; %'# fix for SO formatting - Jonas

%# rows2mergeFrom are the rows where the second row equals the first row of col 1
%# and where the third column is 1. rows2mergeInto indicates the rows from which  the
%# values of the following rows should be added
rows2mergeFrom = find(in(2:end,1) == in(1:end-1,1) & in(2:end,3) == 1) + 1;

out = in;
out(rows2mergeFrom-1,2) = out(rows2mergeFrom-1,2) + out(rows2mergeFrom,2);

%# data that has been shifted up gets subtracted from the 'rows2mergeFrom'-rows
out(rows2mergeFrom,2) = out(rows2mergeFrom,2) - in(rows2mergeFrom,2);

%# remove the ones in the third column from any row that has been involved in a 
%# merge operation
out([rows2mergeFrom;rows2mergeFrom-1],3) = 0

fid = fopen('newData.txt','wt');  
format short g;
fprintf(fid,'%g\t %g\t %g\n',out);  %'# Write the data to the file
fclose(fid);  
A: 

There is no need for an if-statement. What you need is a logical array that allows find to extract the row indices of the rows that are to be merged.

Updated to conform with the new rules

in = [20  10  1
20  11  1
20  15  1
23  10  1
23  10  1
23  12  0];

%# rows2mergeFrom are the rows where the second row equals the first row of col 1
%# and where the third column is 1. rows2mergeInto indicates the rows from which  the
%# values of the following rows should be added
rows2mergeFrom = find(in(2:end,1) == in(1:end-1,1) & in(2:end,3) == 1) + 1;

out = in;
out(rows2mergeFrom-1,2) = out(rows2mergeFrom-1,2) + out(rows2mergeFrom,2);

%# data that has been shifted up gets subtracted from the 'rows2mergeFrom'-rows
out(rows2mergeFrom,2) = out(rows2mergeFrom,2) - in(rows2mergeFrom,2);

%# remove the ones in the third column from any row that has been involved in a 
%# merge operation
out([rows2mergeFrom;rows2mergeFrom-1],3) = 0

out =
20    21     0
20    15     0
20     0     0
23    20     0
23     0     0
23    12     0
Jonas
A: 

As far as I can tell, you don't need to use find or if. You only need logical indexing.

If I understand your question correctly, Amro is right in pointing out that your output doesn't match your description of the logic. Try the following script:

m = [20  10  0; ...
     20  11  0; ...
     20  15  1; ...
     23  10  1; ...
     23  10  1];
merge = (m(1:end-1, 1) == m(2:end, 1)) & (m(2:end,3) == 1)
mnew = m(2:end, :);
madd = m(1:end-1,2) + m(2:end,2);
mnew(merge, 2) = madd(merge);
mnew = [m(1,:); mnew]
Reinderien

related questions