The following syntax does the job on the sample data you provided. First, let's recreate the sample data to demonstrate on:
Data list list/nr (f1) brand (a30).
begin data
1 "GREEN CARE PROFESSIONAL"
1 "GREEN CARE PROFESSIONAL"
1 "GREEN CARE PROFESSIONAL"
2 "HENKEL"
3 "HENKEL"
3 "HENKEL"
3 "HENKEL"
3 "VIZIR"
4 "BIEDRONKA"
4 "BOBINI"
4 "BOBINI"
4 "BOBINI"
4 "BOBINI"
4 "BOBINI"
4 "HENKEL"
5 "VIZIR"
6 "HENKEL"
end data.
dataset name originalDataset.
Now for the restructure.
sort cases by nr brand.
* creating an index to enumerate cases for each combination of `nr` and `brand`.
* This is necessary for the `casestovars` command to work later.
compute ind=1.
if $casenum>1 and lag(nr)=nr and lag(brand)=brand ind=lag(ind)+1.
exe.
* variable names can't have spaces in them, so changing the category names accordingly.
compute brand=replace(rtrim(brand),"","_").
sort cases by nr ind brand.
compute exist=1.
casestovars /id=nr ind /index= brand/autofix=no.