match index of arrays

Hi, I have a problem with matching and indexing. The problem was raised in this forum earlier and was solved. This is how it was raised

A = [9 10  11 12 13 14 15 16 17]'
B = [10 13 16; 999 900 950]'

The desired solution is

M =

       9     0
      10   999
      11     0
      12     0
      13   900
      14     0
      15     0
      16   950
      17     0

This can be attained as follows

M = A*[1 0]
M(ismember(A,B(:,1)),2) = B(:,2)

My problem is exactly the same but the solution is not working. The only difference I have is the size of A and B

A is too large and is of size(A)

 ans =
     1035593           1

B is small and is of size(B)

ans =
    72     2

Column 1 of B contains unique ids that match unique ids in A. Column 2 is the data to be matched given ids.

When I apply the solution above, I get Subscripted assignment dimension mismatch. I am not sure why this is happening, could you kindly advise or provide an alternative solution.

[Stephen23]

>> M = A(:);
>> M(:,2) = 0;
>> [~,ia,ib] = intersect(A,B(:,1));
>> M(ia,2) = B(ib,2)
M =
     9     0
    10   999
    11     0
    12     0
    13   900
    14     0
    15     0
    16   950
    17     0

If you are having problems than you will need to check that every value in column one of B exists once in columns one of A.

Answer 1

[Walter Roberson]

You would get that error if there are columns in B that are not matched in A.

[~, ia, ib] = intersect(A, B(:,1));
M(ia,:) = B(ib,:);

[Tiina]

Thank you. M will be different size to A, I would like to preserve the size of M as A and match rows according to ids. If I pre-allocate M=0xA, then I get a non-singleton dimension error. Is that possible?

[Walter Roberson]

You can start with

M = A*[1 0];

and follow with what I show

[~, ia, ib] = intersect(A, B(:,1));
M(ia,:) = B(ib,:);

Question though: is it possible that your A entries might have duplicated values in column 1 that appear in B ? The code I posted does not expect any matches to be duplicated.

[Tiina]

No, all elements in columns in A and B(:,1) are unique. I checked with unique function.

The remaining problem is to keep the size of M as A in the output. If I preallocate to retain the size of M = A*[1 0] I get a non-singleton error and therefore no solution. If I run your proposed solution, I can match by A and B(:,1) and allocate corresponding elements of B(:,2) to M, however, I do not get the size of M to be equal to the size of A.

[Walter Roberson]

Please attach data that is giving you the problem.

[Tiina]

Please find attached. Thanks.

[Walter Roberson]

I loaded your data and then did

M = A*[1 0];
[~, ia, ib] = intersect(A, B(:,1));
M(ia,:) = B(ib,:);

there were no error messages.

[Tiina]

I restarted Matlab and it works fine. Thank you.