Storing values from inside the third loop to a variable outside the loops

Hi all,

Is it possible to store values from the inside of a third for loop into a variable which is outside the loops? For instance I have this code:

a = 2:1:5;
b = [1 1.2 2 2.3 3 3.3 4 5];

final_data = [];
for ii=1:length(a)
    c = rand(3,8);
    for jj=1:length(b)
        c_data = c(:,jj);
        
        if (a(ii) > b(jj) )
            for xx = 1:length(c_data)
                final_data(xx,ii) = c_data(xx);
            end
        end
    end
end

and I want the values in 'c_data' to be stored inside 'final_data' as rows (the columns are determined by the value of ' ii ' ) for every 'jj' value. But it so happens that the values get over-written for every ' jj ' loop iteration. Or in other words, I was looking for a way to vertically append c_data to final_data for every jj loop iteration without being overwritten.

Thanks

[Matt J]

Could you give an example, with smaller input data dimensions, what the final result is supposed to look like?

[Mathan]

Sure. Let us suppose:

a = a = 1:1:3;

b = 0.1:0.1:0.5;

c = [100 200 300 400 500; 101 201 301 401 501; 102 202 302 402 502; 103 203 303 403 503; 104 204 304 404 504];

I was looking for an output as follows:

Final_data = [100 100 100

101 101 101

102 102 102

.. .. ..

.. .. ..

300 300 300

301 301 301

302 302 302

.. .. ..

.. .. ..

500 500 500

501 501 501

502 502 502

503 503 503

504 504 504 ]

[Voss]

@Mathan Like this?

a = 1:1:3;
c = [100 200 300 400 500; 101 201 301 401 501; 102 202 302 402 502; 103 203 303 403 503; 104 204 304 404 504];

final_data = repmat(c(:),1,numel(a));

disp(final_data)

   100   100   100
   101   101   101
   102   102   102
   103   103   103
   104   104   104
   200   200   200
   201   201   201
   202   202   202
   203   203   203
   204   204   204
   300   300   300
   301   301   301
   302   302   302
   303   303   303
   304   304   304
   400   400   400
   401   401   401
   402   402   402
   403   403   403
   404   404   404
   500   500   500
   501   501   501
   502   502   502
   503   503   503
   504   504   504

[Mathan]

Hi,

Well actually no - sorry if my second example was confusing. I have modified the original question now with a more suitable example which shows that the c matrix changes its value during the loop run and is a random matrix, so infact I really cannot use the repmat(). What I was looking for was a way to append the values from c matrix to final_data as rows without overriding the values.

Answer 1

[Voss]

Maybe something like this?

a = 2:1:5;
b = [1 1.2 2 2.3 3 3.3 4 5];

na = numel(a);
nb = numel(b);
nc = 3;
final_data = NaN(nb*nc,na);

for ii=1:na
    c = rand(nc,nb) % showing c on command-line for reference
    for jj=1:nb
        if (a(ii) > b(jj))
            final_data((jj-1)*nc+(1:nc),ii) = c(:,jj);
        end
    end
end

c = 3×8

0.7756 0.0228 0.8194 0.5967 0.9381 0.2879 0.0092 0.5636 0.6957 0.2733 0.8883 0.8130 0.1560 0.2556 0.6728 0.6247 0.6122 0.4209 0.3791 0.1853 0.5323 0.5289 0.8371 0.9574

c = 3×8

0.3283 0.4400 0.7155 0.8591 0.2053 0.6418 0.8155 0.1471 0.0725 0.2098 0.3402 0.0290 0.3501 0.8085 0.3803 0.9335 0.2055 0.6950 0.4742 0.4666 0.6381 0.4812 0.3803 0.6402

c = 3×8

0.0360 0.5393 0.2289 0.1805 0.0598 0.8810 0.8914 0.7146 0.7247 0.0987 0.1058 0.2307 0.6649 0.0273 0.8709 0.4841 0.1058 0.6794 0.2373 0.1903 0.3412 0.3625 0.4358 0.4167

c = 3×8

0.2973 0.0796 0.7687 0.6992 0.1326 0.6452 0.0892 0.1665 0.4848 0.9148 0.6581 0.6309 0.0303 0.6834 0.3425 0.4914 0.7989 0.3075 0.1044 0.1785 0.6989 0.8545 0.1679 0.0347

disp(final_data)

    0.7756    0.3283    0.0360    0.2973
    0.6957    0.0725    0.7247    0.4848
    0.6122    0.2055    0.1058    0.7989
    0.0228    0.4400    0.5393    0.0796
    0.2733    0.2098    0.0987    0.9148
    0.4209    0.6950    0.6794    0.3075
       NaN    0.7155    0.2289    0.7687
       NaN    0.3402    0.1058    0.6581
       NaN    0.4742    0.2373    0.1044
       NaN    0.8591    0.1805    0.6992
       NaN    0.0290    0.2307    0.6309
       NaN    0.4666    0.1903    0.1785
       NaN       NaN    0.0598    0.1326
       NaN       NaN    0.6649    0.0303
       NaN       NaN    0.3412    0.6989
       NaN       NaN    0.8810    0.6452
       NaN       NaN    0.0273    0.6834
       NaN       NaN    0.3625    0.8545
       NaN       NaN       NaN    0.0892
       NaN       NaN       NaN    0.3425
       NaN       NaN       NaN    0.1679
       NaN       NaN       NaN       NaN
       NaN       NaN       NaN       NaN
       NaN       NaN       NaN       NaN

[Voss]

But the non-NaN values in a column might be discontiguous if a or b is not monotonic:

a = 2:1:5;
% b = [1 1.2 2 2.3 3 3.3 4 5];
b = [1 1.2 4 5 3 3.3 2 2.3];

na = numel(a);
nb = numel(b);
nc = 3;
final_data = NaN(nb*nc,na);

for ii=1:na
    c = rand(nc,nb) % showing c on command-line for reference
    for jj=1:nb
        if (a(ii) > b(jj))
            final_data((jj-1)*nc+(1:nc),ii) = c(:,jj);
        end
    end
end

c = 3×8

0.1083 0.2044 0.3773 0.6838 0.2006 0.4139 0.3873 0.1343 0.6272 0.4463 0.5049 0.1458 0.0585 0.8073 0.8747 0.1989 0.6327 0.2310 0.2846 0.7523 0.9923 0.8388 0.5630 0.1548

c = 3×8

0.7268 0.5949 0.4476 0.5224 0.5686 0.7906 0.6267 0.0658 0.8050 0.6922 0.6834 0.3074 0.3411 0.7303 0.9056 0.8885 0.5745 0.1992 0.9910 0.5114 0.7064 0.4061 0.3204 0.6935

c = 3×8

0.8228 0.8223 0.7012 0.3833 0.8168 0.3479 0.2314 0.5737 0.1215 0.1456 0.9210 0.2621 0.1326 0.6175 0.6664 0.4939 0.5140 0.5830 0.2404 0.9799 0.5821 0.3683 0.2856 0.4276

c = 3×8

0.8164 0.1743 0.8823 0.3160 0.9020 0.0274 0.0497 0.9581 0.1325 0.6394 0.3810 0.0396 0.7746 0.1877 0.7071 0.1677 0.6500 0.9896 0.8840 0.5821 0.9976 0.8418 0.4942 0.0974

disp(final_data)

    0.1083    0.7268    0.8228    0.8164
    0.6272    0.8050    0.1215    0.1325
    0.6327    0.5745    0.5140    0.6500
    0.2044    0.5949    0.8223    0.1743
    0.4463    0.6922    0.1456    0.6394
    0.2310    0.1992    0.5830    0.9896
       NaN       NaN       NaN    0.8823
       NaN       NaN       NaN    0.3810
       NaN       NaN       NaN    0.8840
       NaN       NaN       NaN       NaN
       NaN       NaN       NaN       NaN
       NaN       NaN       NaN       NaN
       NaN       NaN    0.8168    0.9020
       NaN       NaN    0.1326    0.7746
       NaN       NaN    0.5821    0.9976
       NaN       NaN    0.3479    0.0274
       NaN       NaN    0.6175    0.1877
       NaN       NaN    0.3683    0.8418
       NaN    0.6267    0.2314    0.0497
       NaN    0.9056    0.6664    0.7071
       NaN    0.3204    0.2856    0.4942
       NaN    0.0658    0.5737    0.9581
       NaN    0.8885    0.4939    0.1677
       NaN    0.6935    0.4276    0.0974

And I guess you'd want those non-NaNs at the bottom of columns 2-4 to be moved up?

[Mathan]

oh yes this was something which I was looking for! Thank you - well I would be fine with intermediate nan values in columns 2-4 since that would serve my purpose.

Thanks again

[Mathan]

But it would be great if I could remove the leading NaNs and shift everything upwards - like for example if the 2nd column had begun with 3 NaNs and then followed by 0.3283, 0.0725, etc then is it possible to remove those leading NaNs so that my second column now begins with 0.3283, 0.0725, etc?

[Voss]

You can avoid having any NaNs above any non-NaN c values by keeping track of how many rows have been written to in each column:

a = 2:1:5;
% b = [1 1.2 2 2.3 3 3.3 4 5];
b = [1 1.2 4 5 3 3.3 2 2.3];

na = numel(a);
nb = numel(b);
nc = 3;

final_data = NaN(nb*nc,na);
last_row = zeros(1,na);

for ii=1:na
    c = rand(nc,nb) % showing c on command-line for reference
    for jj=1:nb
        if (a(ii) > b(jj))
            final_data(last_row(ii)+(1:nc),ii) = c(:,jj);
            last_row(ii) = last_row(ii)+nc;
        end
    end
end

c = 3×8

0.8679 0.1344 0.7409 0.8991 0.6924 0.2131 0.9353 0.1963 0.1268 0.5915 0.1404 0.2434 0.2801 0.9089 0.8597 0.3815 0.4814 0.8942 0.4151 0.6657 0.6755 0.9031 0.7325 0.5550

c = 3×8

0.0562 0.0343 0.9638 0.9161 0.0169 0.6211 0.6252 0.8444 0.2833 0.2906 0.9503 0.5871 0.1685 0.9125 0.8346 0.3167 0.5431 0.7185 0.8937 0.3508 0.7467 0.8985 0.6085 0.8455

c = 3×8

0.2012 0.3699 0.6199 0.4750 0.9173 0.8522 0.9058 0.6655 0.4513 0.2709 0.4633 0.5649 0.7279 0.1773 0.4038 0.2587 0.3472 0.4188 0.9591 0.1697 0.4908 0.2343 0.5552 0.5674

c = 3×8

0.4964 0.4386 0.9688 0.5840 0.5509 0.1135 0.8336 0.3020 0.8426 0.6906 0.7833 0.6806 0.3329 0.6814 0.2129 0.4306 0.3220 0.0782 0.9554 0.0162 0.3916 0.7909 0.6537 0.9954

disp(final_data);

    0.8679    0.0562    0.2012    0.4964
    0.1268    0.2833    0.4513    0.8426
    0.4814    0.5431    0.3472    0.3220
    0.1344    0.0343    0.3699    0.4386
    0.5915    0.2906    0.2709    0.6906
    0.8942    0.7185    0.4188    0.0782
       NaN    0.6252    0.9173    0.9688
       NaN    0.8346    0.7279    0.7833
       NaN    0.6085    0.4908    0.9554
       NaN    0.8444    0.8522    0.5509
       NaN    0.3167    0.1773    0.3329
       NaN    0.8455    0.2343    0.3916
       NaN       NaN    0.9058    0.1135
       NaN       NaN    0.4038    0.6814
       NaN       NaN    0.5552    0.7909
       NaN       NaN    0.6655    0.8336
       NaN       NaN    0.2587    0.2129
       NaN       NaN    0.5674    0.6537
       NaN       NaN       NaN    0.3020
       NaN       NaN       NaN    0.4306
       NaN       NaN       NaN    0.9954
       NaN       NaN       NaN       NaN
       NaN       NaN       NaN       NaN
       NaN       NaN       NaN       NaN

[Mathan]

Thank you.