# assign numbers to observations

7 views (last 30 days)
matla6123 on 5 Apr 2017
Commented: Stephen Cobeldick on 10 Apr 2017
Hi I have a large dataset involving observations of names in a cell like below.
ABC1
AB56
AC46
AC90
I have another dataset "values" that is made up of doubles, with observations for each of the names listed above.
3x8 doubles
3x8 doubles
6x8 doubles
5x8 doubles
The datasets are specifically ordered so that the third observation in "values" has the third name. I want to give each name a number referring to its position in the original vector. E.g. AC90 is the fourth in the list so everywhere AC90 appears, the number 4 is listed.

#### 1 Comment

Stephen Cobeldick on 10 Apr 2017
@matla6123: whatever you do, do not try to create/access variable names dynamically. Dynamically accessing variables names is slow, buggy, and obfuscated way to write code:
Much better would be to use a non-scalar structure, as Jan Simon has shown you.

Jan on 10 Apr 2017
If the names, values and their order should be represented, store them e.g. in a struct array:
S(1).Name = 'ABC1';
S(1).Data = <3x8 doubles>
S(2).Name = 'AB56';
S(2).Data = <3x8 doubles>
and so on. Now S(k) contains the k's data set with its name.
Cells are an alternative:
NameList = {'ABC1', 'AB56', 'AC46', 'AC90'}
DataList = {3x8 doubles, 3x8 doubles, 6x8 doubles, 5x8 doubles}
Now the name NameList[k} corresponds to the data DataList{k}. As far as I understand, you have this data format already.
Trying to create a variable, which is called 'AC90' would be a very bad idea, see Answers: Don't EVAL!.
Another idea would using the names as fieldnames:
S.ABC1 = <3x8 doubles>
S.AB56 = <3x8 doubles>
...
Then:
key = 'AB56';
S.(key)