how to apply condition on multiple columns?

I have a large data set that is unequal in every way so to get things done i have to apply condition on the element of the specific columns. E.g

1 1 234
1 1 234
1 1 180
1 2 345
1 2 674
.
.
.
8  1  453
8  1  678
8  2  478
8  2  345
8  2  345
.
.
.
31 21 456
31 22 452
31 22 345
31 23 670
31 23 567
31 23 456

what i want the programme to do is that check the value in column 1 if it equals 1 the check the value in column 2 if it equals 1 then take the mean of all the values in column 3 corresponding to column 2. similarly i want it to repeat for all the values in column 2 for 1st then all the other numbers in column 1 till last number in this case which is 31.

[Stephen23]

See my answer for a neater and more efficient solution than the accepted answer.

Answer 1

[KSSV]

A = [1 1 234
1 1 234
1 1 180
1 2 345
1 2 674
8  1  453
8  1  678
8  2  478
8  2  345
8  2  345
31 21 456
31 22 452
31 22 345
31 23 670
31 23 567
31 23 456] ;

B = [1 1] ;

[idx,val] = ismember(A(:,1:2),B,'rows') ;
% get mean
iwant = sum(A(idx,3))/nnz(idx)

[Mahrukh Farooq]

It worked WOW! i just had to put it in a loop.

for i=1:31;
 for ii=1:24;
 B = [i ii] ;

[idx,val] = ismember(te(:,2:3),B,'rows') ;
% get mean
iwant(i,ii)= sum(te(idx,5))/nnz(idx);
end
end

Thanks a lot.GOD bless you.

"No one is useless in this world who lightens the burdens of the others"

~Charles Dickens.

[Stephen23]

See my answer for neater and more efficient solution without the inefficient loops and multiple ismember calls.

Answer 2

[Stephen23]

This can be done very efficiently with just one call to accumarray, no ugly loops are required:

 >> out = accumarray(A(:,1:2),A(:,3),[],@mean);

And checking:

>> out(1,2) % mean when A(:,1)==1 & A(:,2)==2
ans =  509.50
>> out(8,1) % mean when A(:,1)==8 & A(:,2)==1
ans =  565.50

[Andrei Bobrov]

+1