Creating multiple variables at once

[Dan Palmer]

I might be losing my mind, but I thought I once ran across a component of Stata syntax (command, extension, option, etc.) that could do the following using only one line of code:

In this scenario, I have the following:

1. a string grouping variable (called group) that takes two possible values: a or b
2. a numeric outcome variable (called score) that takes values 1 to 100

If the syntax I'm thinking of actually exists, it could generate two new variables at once:

1. score_a would take values of score when group==yes
2. score_b would take values of score when group==no

Does this sound remotely familiar to anyone? Or am I inventing this? After significant web searching this morning I haven't been able to find anything like what I have in mind. But I'd swear I encountered something like this at some point in the past.

If anyone could put me out of my misery one way or another, I'd be obliged.

[George Ford]

you want the mean?

the data is there already, so not obvious what you're after.

just to create new variables,

g score_a = score if group=="a"
g score_b = score if group=="b"

if you want them to have 0 values rather than missing, you could use cond() or replace if ==.

[Dan Palmer]

I want the original raw values from the score column, not a calculated value derived from them. Your suggested two lines of code do what I want, but I'm wondering if there is a Stata command that achieves the same purpose using only one line.

[Nick Cox]

help separate

seems to be what you're seeking.

[Dan Palmer]

That's it! Thanks Nick!

I should have asked here much earlier; that command really resists being discovered through Googling.

[Nick Cox]

As the original author, the command name seems natural to me.

split was taken at about the same time for something different (but not completely different).