CODEWITHSUNDEEP
matlabmatrixvector
Masan
Masan

Reputation: 345

Vectorizing Computation from Table Data in Matlab

I have two tables for some values and want to do a matrix calculation with it by Matlab. I know how to do it in scalar and then make a vector by combining these calculated scalar values. The problem description is as follows: 

Table 1  
in|out|  IMP  |  Ang    
1 | 2 | 0.871 | 0.281
1 | 4 | 0.304 | 0.304
1 | 5 | 0.064 | 0.064
2 | 3 | 0.108 | 0.108
3 | 4 | 0.297 | 0.297
4 | 5 | 0.297 | 0.297
%%%%%%%%%%%%%%%%
Table 2
indx|     val  | shift
  1 |   1.0003 | 0.1
  2 |   0.9784 | 0.2
  3 |   1.0019 | 0.09
  4 |   1.0039 | 0.1
  5 |   1.0475 | 0.5

The scalar values are calculated this way. We must get 6 scaler values based 'in' and 'out' at table one. these values are C12,C14,C15,C23,C34,C45, other elements like C35 is 0. where 

C12=(1.003/0.871)*(cos(0.281)-0.978*cos(0.281-0.2+0.1)).

In other words, 

C12=(val(1)/IMP(1,2))*(cos(Ang(1,2))-val(2)*cos(Ang(1,2)-shift(2)+shift(1))).

I tried to make a matrix using this table. for example I've constructed IMP_mat for 3rd column of table 1 (IMP_mat(1,2)=0.871 and so on). But in equations for calculating C12,.. I am getting error as some values for IMP_mat are zero as they are not defined in table (for example IMP_mat(1,3)=0).
%% EDIT:
Based on @krisdestruction's suggestion, I have written the code in table format and converted it in matrix form. But I don't know what to do next.

in=[1 1 1 2 3 4]';
out=[2 4 5 3 4 5]';
IMP=[0.871 0.304 0.064 0.108 0.297 0.297 ]';
Ang=[0.281  0.304   0.064   0.108   0.297   0.297]';
indx=[ 1    2   3   4   5]';
val=[1.0003 0.9784  1.0019  1.0039  1.0475]';
shift=[0.1  0.2 0.09    0.1 0.5]';
T1=table(in,out,IMP,Ang)
T2=table(indx,val,shift)
M1 = table2array( T1 );
M2 = table2array( T2 );

And now, I think the problem description for c12,... becomes more complex. Also, it seems using this table thing expend a lot of compution. Is that right? My actual data have thousands of elements.
%% My Attemp:
I have tried to write them manually in matrix forms and do the calculation. But still I don't know how to vectorize my relationship for C. Here is the script:

  %IMP                             %Ang
 IMPmat( 1  ,  2 )=   0.871  ;    Angmat( 1 ,   2 )=    0.281   ; 
 IMPmat( 1   , 4 )=   0.304  ;    Angmat( 1 ,   4 )=    0.304   ;
 IMPmat( 1  ,  5 )=   0.064  ;    Angmat( 1 ,   5 )=    0.064   ;
 IMPmat( 2  ,  3 )=   0.108  ;    Angmat( 2 ,   3 )=    0.108   ;
 IMPmat( 3   , 4 )=   0.297  ;    Angmat( 3 ,   4 )=    0.297   ;
 IMPmat( 4  ,  5 )=   0.297  ;    Angmat( 4 ,   5 )=    0.297   ;

% 
val=[1.0003 0.9784  1.0019  1.0039  1.0475]';
shift=[0.1  0.2 0.09    0.1 0.5]';
%%
C12=(val(1)/IMPmat(1,2))*(cos(Angmat(1,2))-val(2)*cos(Angmat(1,2)-shift(2)+shift(1)))
C14=(val(1)/IMPmat(1,4))*(cos(Angmat(1,4))-val(4)*cos(Angmat(1,4)-shift(4)+shift(1)))

%% EDIT(New)
I have tried something else. It gives some results without error. But I am not sure if I calculated correctly. But for the first case, it gives the correct result. Here is my Attempt: 

  in= [1 1 1 2 3 4]';
  out=[2 4 5 3 4 5]';
  IMP=[0.871    0.304   0.064   0.108   0.297   0.297 ]';
  Ang=[0.281    0.304   0.064   0.108   0.297   0.297]';
  indx=[ 1  2   3   4   5]';
  val=[1.0003   0.9784  1.0019  1.0039  1.0475]';
  shift=[0.1    0.2 0.09    0.1 0.5]';
  C=(val(in)./IMP).*(cos(Ang)-val(out).*cos(Ang-shift(out)+shift(in)))

Upvotes: 0

Views: 135

Answers (1)

krisdestruction
krisdestruction

Reputation: 1960

It looks like you're trying to do specific operations that can be easily done/vectorized in a matrix. In that case, I would recommend you convert it into a matrix using table2mat first and perform your calculation. It's easy because you have all numeric data.

Suppose you have a table formatted like this. 

Var1    Var2    Var3    Var4
____    ____    ____    ____
1       2       2       1   
1       4       5       5   
1       5       1       3   
2       3       3       7   
3       4       5       5   
4       5       5       5

The code to generate is below.

T = table([1 1 1 2 3 4]',[2 4 5 3 4 5]',[2 5 1 3 5 5]',[1 5 3 7 5 5]')

You can use this line of code to convert it to a matrix array.

M = table2array( T );

You can then proceed to perform your calculations using matrices and for each of the C12,C14,C15,C23,C34,C45 in vector form (if possible).

Then you can use the following line of code to do your calculation.

C = val(in) ./ IMP .* cos( Ang(out) ) - val(out) .* cos( Ang(in) - shift(out) + shift(in) );

The result is shown here.

>> C
C =
    2.1772
    8.2435
    2.1805
   14.3389
    8.2204
    2.2290

Upvotes: 1

Related Questions