how to generate columns from comma separated string variables

[Janet Yang]

Hello,

I have responses from a questionnaire, where several questions have "select all that apply." These questions have been entered into STATA as a single string variable, separated by commas.

Is there a way to generate new columns for each unique response in the cell, and have the values appear in their specific newly generated columns?

For example, we have a list of activities numbered 1-11. A single cell may contain any combination of these numbers, such as (1,2,5,7,9).

We tried the following command: "split activities, parse(,) gen(activity)" and it produces columns titled Activity1 through Activity 11, but all numbers in the columns in order (ie. 1 will be in activity1, 2 will be in activity2, but then 5 will appear in activity3 etc. rather than under activity5).

Is there a way to do this? Thanks in advance for any advice.

[Roberto Ferrer]

I think you want something like this:

clear
set more off

input ///
str25 answer
"1,2,5,7,9"
"5,3,9,2"
end

list

split answer, parse(,) destring gen(activity)
gen id = _n

reshape long activity, i(id) j(act)
drop act answer
drop if missing(activity)

gen act = 1
reshape wide act, i(id) j(activity)

list

(Data in long form is usually more useful than it is in wide form.)

[Jorrit Gosens]

Alternatively, see this discussion here: http://www.statalist.org/forums/foru...rds-in-phrases

[Aga Smith]

I am getting an error after " reshape long activity, i(id) j(act)"
(note: j = 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15)
I/O error writing .dta file
Usually such I/O errors are caused by the disk or file system being full.
r(693);

May I ask for an advice how can I deal with it?

Thank you.

I think you want something like this:

clear
set more off

input ///
str25 answer
"1,2,5,7,9"
"5,3,9,2"
end

list

split answer, parse(,) destring gen(activity)
gen id = _n

reshape long activity, i(id) j(act)
drop act answer
drop if missing(activity)

gen act = 1
reshape wide act, i(id) j(activity)

list

(Data in long form is usually more useful than it is in wide form.)