Select a subsample from a database

[Adolfo Aramburu]

Hi, I have a database with 5000 obs of which I only need 600. I have used the following syntax:

set seed 10101
sample 600 if valid == 1, count (valid is a conditional variable for the selection)

However, running 'sample' removes all other observations. I want to create a variable that puts 1 to the selected observations and 0 to the unselected ones, instead of deleting them.

Any advice for it?

Thank you,
Adolfo Aramburu

[Ali Atia]

The randomselect command might be what you're looking for. You can install it using

ssc install randomselect

Then, you can generate the dummy sample variable using this command:

randomselect if valid == 1, gen(sample) n(600)

[Adolfo Aramburu]

It worked perfect, thanks