Adding string values separated by comma as new variables

[Federico Nutarelli]

Hi all,

as the title suggests I would like to add as new observations string values included in a string and separated by a comma.
So, for instance if I have the following:

* Example generated by -dataex-. To install: ssc install dataex
clear
input obs str114 atc3_ephmra
1 "A10H, S1X"
2 "A10X, C1D, J2A, L4X, D5B"
3 "B10X, F1A, J2C, S4B, K5D"

I would like to end up with:

input obs str114 atc3_ephmra
1 A10H
2 S1X
3 A10X
4 C1D
5 J2A
6 L4X
7 D5B
8 B10X
9 F1A
10 J2C
11 S4B
12 K5D

I guess thee first step is to split at the comma if stripos(atc3_ephmra,","), but I then am unable to go on.
Any help is appreciated.

Thank you all.

[Nick Cox]

Thanks for the data example.

* Example generated by -dataex-. To install: ssc install dataex
clear
input obs str114 atc3_ephmra
1 "A10H, S1X"
2 "A10X, C1D, J2A, L4X, D5B"
3 "B10X, F1A, J2C, S4B, K5D"
end 

local y atc3_ephmra 
split `y', parse(,)
drop `y'

reshape long `y', i(obs) j(which)
drop if missing(`y')

list

[Federico Nutarelli]

Nick Cox Thank you a lot for the reply. It works perfectly. Just a question: does it work also if there are repeated observations in the data?
E.g.

 * Example generated by -dataex-. To install: ssc install dataex clear input obs str114 atc3_ephmra
1 "A10H, S1X"
2 "A10X, C1D, J2A, L4X, D5B"
3 "A10X, C1D, J2A, L4X, D5B"
4 "A10X, C1D, J2A, L4X, D5B"
5 "B10X, F1A, J2C, S4B, K5D"
6 "B10X, F1A, J2C, S4B, K5D"
7 "B10X, F1A, J2C, S4B, K5D" end

Theoretically it should since it is a reshape long using as stubs the splitter values of atc3_ephmra right?

[Nick Cox]

It is easy enough to experiment and find out, but yes, repeated values are not a problem there. Or, in principle, nothing in the code drops duplicates (except duplicate missings). Sometimes you want that, but it's an extra step. .

[Federico Nutarelli]

Thank you very much for the help!