# index of zero prolem in matlab

Asked by huda nawaf on 29 Dec 2011
Latest activity Commented on by Walter Roberson on 29 Dec 2011

hi, is there solution for index of zero problem. I'm facing this problem in this code I'm working to design code of local sequence alignment(smith_watrman algorithm)

it is very necessary to fill the first column (zero column) and the first row(zero row) with zero value as initial matrix, then fill the other cells of array with values gradually.

how can run this code? is there another way?

for j=1:length(x)

`    mat(0,j).scor=0;end;`

for i=1:length(y)

`    mat(i,0).scor=0;end;`

%%%%fill matrix

` for i=1:length(y)`
`     for j=1:length(x)`
`     %%%%%% calculate match score`
`     letter1=x(j-1:j-1);letter2=y(i-1:i-1);`
`     etc....`

end;end;end;

thanks

## Products

No products are associated with this question.

Answer by Andrew Newell on 29 Dec 2011

You seem to have a structure array. You could initialize it using

```mat = repmat(struct('scor',0),length(y),length(x));
```

## 1 Comment

Walter Roberson on 29 Dec 2011

Yes, that should be effective enough in most cases.

This will, however, use the "same" scalar 0 to initialize all of the locations, with new memory allocated the first time it becomes necessary to write a value at a location. In some cases it can be important to allocate all of the memory first. The vectorized ways of doing that tend to temporarily require about double the memory, though, so a loop can end up being what you should use.

It will not be possible to assign to index 0 -- not unless you create your own OOP class with a subsref and subsassgn method that understands an index of 0. If there was a good reason why that was necessary (instead of just changing the index calculations) then some of the contributions in the File Exchange might help.