How total of column of a matrix can be kept be same????

If I have a matrix A A=[ 0.4000 0.6000 0.8000 0.8000 1.0000 1.6000 0.9333 1.2333 0.0000 0.0000; 1.2000 0.6000 0.8000 0.8000 1.0000 0.4000 0.5333 0 0 0; 1.2000 1.4000 1.6000 1.6000 1.8000 2.4000 2.5333 2.8333 4.0000 4.0000; 1.2000 1.4000 0.8000 0.8000 0.2000 0 0 0 0 0] I want to keep the total of column to remain same as here the total of each column is as:

sum(A(:,1))

ans =

     4

>> sum(A(:,2))

ans =

     4

>> sum(A(:,3))

ans =

     4

>> sum(A(:,4))

ans =

     4

>> sum(A(:,5))

ans =

     4

>> sum(A(:,6))

ans =

    4.4000

>> sum(A(:,7))

ans =

     4

>> sum(A(:,8))

ans =

    4.0667

>> sum(A(:,9))

ans =

     4

>> sum(A(:,10))

ans =

     4

at 6th and 8th column its becoming more than 4 .... so here how can round of this to 4??? plz plz help.....

[David Sanchez]

One doubt: Do you want to modify the value of any of the column elements in order to have a column such that sum = 4?

[suchismita]

yes....if needed some should be deleted to get 4

[suchismita]

here in my example matrix can you help me in modifying without disturbing the zeros...please...like if in a column sum is more than 4 it should minus the rest and if its less it should add to make it 4... can you help me in that way...

Answer 1

[Jos (10584)]

This is called a normalisation procedure.

A = rand(10,6) % some random data
sumA = sum(A,1) % unequal column sums

normfun = @(a,b) 4 * (a/b)
Anorm = bsxfun(normfun, A, sumA) % normalise each column

sum(Anorm,1) % all 4, check!

[suchismita]

thank u so much... but here in my example matrix can you help me in modifying without disturbing the zeros...please...like if in a column sum is more than 4 it should minus the rest and if its less it should add to make it 4... can you help me in that way...