How to vectorize this operation

Hi,

Could you help guide how I can vectorize this operation? I have the original fullset matrix which i want to retain and populate for each of the locID (4 locID total). For each locID, it looks up the loss values from ELT when there's a match on eventid.

fullset= struct;

fullset.yearid= [1     1   2   3   4    5     6    7    7    7]';

fullset.eventid=[201 202 203 204 201   210   208  209 210  210]';

fullset.indx   = [1:10]';

fullset= struct2table( fullset);

ELT= struct;
ELT.eventid= [ 201 203 209 210 210 210 201  210]';

ELT.loss   = [ 100 100 100 200 1     1  5   1]';

ELT.locid  = [ 1     1   1   1 2     3  4   4]';

ELT= struct2table( ELT);

This is the operation I used to come up with full matrix set [10x4]. It has the full sequence from fullset, and for each indx, it's filled in with the loss from ELT of the same eventid

uniqLocID= unique( ELT.locid);

nLoc= length( uniqLocID);
simLocFull= deal( zeros( length( fullset.eventid), nLoc) ); 

for runi= 1: nLoc

      tloc = (ELT.locid== uniqLocID( runi)); % (JlocID== runi); 
      tdata= struct;
      [tdata.eventid, tdata.loss]= deal( ELT.eventid(tloc), ELT.loss(tloc));

      [ tlocMaster, locELT]= ismember( fullset.eventid, tdata.eventid);

      simLocFull( tlocMaster, runi)=  tdata.loss( locELT( tlocMaster)); 

end % for runi

I tried using outerjoin, but was unsuccessful.

tall=outerjoin( fullset, ELT,'Type','LEFT','MergeKeys', true, 'LeftKeys',{'eventid'},'RightKeys',{'eventid'}, 'LeftVariables',{'yearid','eventid'},'RightVariables',{'locid','loss'});

Thanks in advance.

Answer 1

[Pete sherer]

Each locid in the ELT has a unique set of eventid by the way