Using intersect to assign a classification

Hi there!

I've created a classification matrix and have used intersect to find the common cell, but the problem is that I want the information in the cell next to it.

Example:

US-Goo  US-Ha1  US-Ha2
GRA  DBF  ENF

What I have:

a=length(landclass)
IGBPclass=intersect(sitename, landclass)
IGBPclass='US-Ha1'

Ideally I want IGBPclass to be the second 3 letter classification.

Thanks!

[Matt J]

Show us sitename and landclass. Show the result you are getting. Show the result that you want. Why do you overwrite IGBPclass in the 3rd line of code,

 IGBPclass='US-Ha1'

[Kate]

The sitename is US-Ha1

The landclass that encapsulates this site is DBF.

I would like the IGBPclass to be DBF, not US-Ha1.

The third line is the output of the intersect function.

I've tried the 'rows' syntax in an attempt to grab the matching data, but get this error message.

"Warning: The 'rows' input is not supported for cell array inputs. > In cell.intersect>cellintersectlegacy at 154 In cell.intersect at 143 "

Thanks for your help!

Answer 1

[Matt J]

The third line is the output of the intersect function.

That's very confusing since we cannot distinguish it from other lines of code. This is how you should show contents of variables to the public,

    >> landclass, sitename

    landclass = 

        'US-Goo'    'US-Ha1'    'US-Ha2'
        'GRA'       'DBF'       'ENF'   

    sitename =

    US-Ha1

If the above is the form your data is in, I suspect you are trying to do this:

    >> logicalIndex=ismember(landclass(1,:),sitename);
    >> IGBPclass=landclass{2,logicalIndex}

    IGBPclass =

    DBF

[Kate]

That worked. Thank you for your help and sorry for being confusing.

Answer 2

[the cyclist]

I don't fully understand what you are trying to do. However, if you read

doc intersect

you will see that you can call that command with more than one output. The second and third outputs are the indices for the elements that are in the intersection of the two sets. Maybe you can use those indices to do what you want.