# array: concatenate two columns of integers to one colum

3 views (last 30 days)
William Sampson on 13 Sep 2012
(Matlab novice)
I thought this would be so easy...
I have a 2-d array of integers and I want to combine two columns like this:
... 123 456 ...
... 123 456 ...
etc...
I want:
... 123456 ...
... 123456 ...
etc...
as single numbers (that is, I want (123 * 1000) + (456)
The number of rows will be variable from one run to the next.
Any help would be appreciated.
Jan on 13 Sep 2012
The question is not clear. What is the wanted output for: [1,2,3,4; 5,6,7,8; 9,10,11,12] ?

Honglei Chen on 13 Sep 2012
Edited: Honglei Chen on 13 Sep 2012
a = randi([1 100],[5 4]) % a 5x4 integer matrix
b = cellfun(@num2str,{a(:,2:3)},'UniformOutput',false)
c = cellfun(@(x) strrep(x,' ',''), cellstr(b{1}),'UniformOutput',false)
a(:,2) = str2double(c)
a(:,3) = []
##### 2 CommentsShowHide 1 older comment
Honglei Chen on 13 Sep 2012
@Jan, thanks for the tip. That's a good one. I think you missed a b in your second expression.

Tom on 13 Sep 2012
Edited: Tom on 13 Sep 2012
A=[1 2 3;3 5 7];
B=[7 8 9;4 5 6];
%
C=[A B];
w=size(C,2);
D=10.^(w-1:-1:0);
Out=sum(bsxfun(@times,D,[A B]),2)

Jan on 13 Sep 2012
Edited: Jan on 13 Sep 2012
A = [1234, 567; 1, 234];
B = A(:, 1) .* nextpow10(A(:, 2)) + A(:, 2);
function Y = nextpow10(X)
Y = 10 .^ floor(log10(abs(n)) + 1); % [EDITED]
Please test this before using, I cannot run Matlab currently.
Andrei Bobrov on 13 Sep 2012
Edited: Andrei Bobrov on 13 Sep 2012
A=randi(67,6,4)
n = ceil(log10(A(:,2:end)));
out = sum(A(:,1:end-1).*(10.^fliplr(cumsum(fliplr(n),2))),2)+ A(:,end)

William Sampson on 13 Sep 2012
Yeah - so
1) I'm not getting it and
2) I seem to not have been clear enough (sorry...)
what I have is a data array with 59 columns and an indeterminate number of rows and I want to combine 2 (for example) columns as described above. In reality it is 5 columns, 8 - 12, of 16 digits each, but I'll start with getting 2 of them done.
So I have, e.g., rawData(:,9) and rawData(:,10) to combine
##### 2 CommentsShowHide 1 older comment
Jan on 13 Sep 2012
Edited: Jan on 13 Sep 2012
See my suggestion:
B = rawData(:, 9) .* nextpow10(rawData(:, 10)) + rawData(:, 10)
result = [rawData(:, 1:8), B, rawData(:, 11:end)]
I do not see, why the suggested method should not work.
Btw. What do you expect as result of two 16 digit numbers? You can store only 16 digits in variables of the type DOUBLE.

Ryan on 13 Sep 2012
Edited: Ryan on 13 Sep 2012
A = [1234, 567; 1, 234]; %Sample Data
% Figure out how many spaces to shift the left column
C = ceil(log10(A(:,2)));
% Find areas where the value is a multiple of 10 or is 1
idx = A(:,2) == 1 | rem(A(:,2),10) == 0;
C(idx) = C(idx) + 1; % Handle the exception
% Determine how much to shift the left column
C = 10.^C;
% Add in the right column
A = A(:,1).*C + A(:,2);

### Categories

Find more on Logical in Help Center and File Exchange

### Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by