Loops vs. Vectorization

How can I replace this:

for i = 1:size(zensus,1)
    disp(i)
    for j = 1:size(spezQ,1)
        if zensus.GBT(i)==spezQ.GBT(j)&&zensus.BJR(i)==spezQ.BJR(j)&&zensus.ZLW(i)==spezQ.ZLW(j)
            zensus.SQ(i) = spezQ.SQ(j);
        end
    end
end

with a vectorized version?

I am trying to assign certain values (spezQ) to combinations of categories in my data (zensus).

Any help would be incredible.

Edit: The data looks like this:

zensus: [205240x6]

 HZT  FLW  ZLW  HHG  GBT  BJR   SQ
 
  1  1  1  1  1  1  ?
 
  1  1  1  2  2  1  ?
 
  1  1  2  1  1  1  ?

spezQ: [27x4]

  GBT  BJR  ZLW  SQ 
 
  1  1  1  212,45325 
 
  1  1  2  192,6525 
 
  1  1  3  183,0135 
 
  1  2  2  161,2365 

For example: zensus.SQ(1) should be equal to spezQ.SQ(1) because of the matching values of GBT, BJR and ZLW.

My loop takes forever because of the length of zensus. So I am looking for faster code!

[Mathieu NOE]

hello

what are you looking for ? faster code ?

maybe if you coud provide an input data file to test it...

[Michael Pietruschka]

I edited my question. Thanks for answering

Answer 1

[Jon]

It may be useful for you to know that for example

M = [3 5 7] == [2; 3; 4;] % row vector == column vector

gives a matrix

3×3 logical array

0 0 0

1 0 0

0 0 0

You can find the matching row and column using indices

[i,j] = find (M)

i =

2

j =

1

So you could probably do something like:

[i,j] = find(zensus.GBT==spezQ.GBT'&&zensus.BJR==spezQ.BJR'&&zensus.ZLW==spezQ.ZLW')
zensus.SQ(i) = spezQ.SQ(j)

I may have mixed up the i's and j's but I think you will get the idea

[Michael Pietruschka]

Thank you for your input! I will try it out tomorrow.